Luma Group

September 2025

Healthcare is one of the most data-rich and capital-intensive sectors yet remains decades behind in analytics. High-stakes decisions rely on incomplete data and systems that are slow, manual, error-prone, and expensive. The result is billion-dollar missteps, extended timelines in an industry already operating on decade-long horizons, and delays in bringing life-saving medicines to patients.

Despite rapid advances in LLMs, these breakthroughs have yet to meaningfully change how strategic decisions are made in biotech. Today’s LLMs are fluent language tools. They can summarize dense papers, extract entities, or polish prose, but language is not the same as evidence-backed strategy.

Biotech demands reasoning across fragmented, multi-modal data: trials, patient outcomes, regulatory precedent, and competitive context, all shifting in real time.

Addressing this call for platforms that are purpose-built for biotech, fluent in the networked language of biology, and capable of turning vast, messy evidence into actionable, trusted insights. Only with tools like these will strategic decisions in biotech and healthcare match the rigor the field demands.

In this whitepaper, we (1) examine how strategic decisions in biotech are made today, (2) show where today’s LLMs help and where they fail, (3) propose a bio-native architecture built on a shared data foundation, multi-hop reasoning, and UX-driven validation loops, and (4) introduce LABI, Luma Group’s AI for Biotech Intelligence, and how we aim to apply these principles in practice.

How Are Strategic Decisions Made in Biotech Today?

Consider a typical diligence scenario to evaluate a new therapeutic candidate. You ask an AI system: “What is the current standard of care in this disease, and how does this candidate compare on endpoints, patient selection, safety, and competitive position? Pull the underlying trials, relevant patient datasets, regulatory precedents, and commercial context, and return it as a citation-backed memo with tables.”

It is a compelling vision: a single click yielding evidence-backed answers that can be trusted. But, when pushed with inquiries like this, today’s LLMs fall short. Results are often incomplete and inaccurate, with hallucinations that make them unfit for high-stakes strategic decisions.

Beyond these limits, the diligence process in biotech is uniquely complex and fundamentally different from generalist investing. It requires deep, cross-functional expertise across domains such as foundational biology, discovery and development, clinical data, regulatory precedent, reimbursement strategy, and commercial dynamics. Each domain is a fragmented and hard-to-reach pool of millions of data points that are still largely hunted down and tracked manually. Connecting them into a coherent view requires months of labor-intensive and error-prone work. Moreover, the landscape is constantly shifting. As the right information surfaces, connections emerge both within and across domains, which then need to be tested against multiple possible outcomes. A failure in a novel class, for example, can cascade across development, commercial potential, and competitor pipelines, inducing a rippling effect across domains.

To address these complexities, the industry still takes a piecemeal approach: scaling by headcount, subscribing to multiple data platforms, augmenting existing workflows with generic AI, and tracking and pulling data into manual workflows. As a result, high-stakes biotech decisions are made with fragmented, incomplete data, through manual, error-prone processes that risk costly missteps.

Figure 1: Biotech decision-making today: High-stakes biotech decisions are made with fragmented, incomplete data, through manual, error-prone processes that risk costly missteps.

AI’s Impact and Limitations in Biotech Decision Making

The real constraints on AI in biotech are not computational power or algorithms but rather data and complexity.

Existing LLM successes hit complexity walls: Even where LLMs work well, they struggle with complexity. Code writing AI has immediate feedback when code compiles or breaks but fails with large codebases. Travel planning LLMs can pull from standardized databases but fail when preferences conflict or constraints shift. These examples show that AI breaks down when reasoning must span interconnected systems, even with built-in advantages like immediate feedback and standardized data.

Biology uniquely amplifies these challenges: Biotech faces the same complexity scaling problems without any structural advantages. The heterogeneity of biological data creates fundamental problems. Clinical trials for identical indications might report results in completely different formats, representing disagreements about what constitutes meaningful measurement in biological systems. Reasoning becomes exponentially more complex when drug interactions, competitive landscapes, protein networks, and patient populations are layered on.

Figure 2: Complexity and network effects in biotech: Biotech complexity emerges from connections: foundational biology to clinical data, regulatory strategy to reimbursement, market dynamics to competitive intelligence. Each layer interlocks, creating numerous reasoning challenges.

User experience solutions remain unexplored: Other domains experiment across the full spectrum from autocomplete to autonomous agents, yet even well-resourced teams are pivoting back to human-in-the-loop systems for complex tasks. Biotech has barely explored beyond basic autocomplete. This represents an opportunity: biotech can learn from experimentation in other domains and design purpose-built solutions for its unique data structures and reasoning requirements.

Building a Bio-Native Intelligence Platform

Building AI systems that can reliably reason across interconnected biotech datasets requires three fundamental design principles:

Get the data in one place, and one language: Before clever prompts, we need a common evidence layer of disparate data pulled into a standard format with clear labels. Today, the facts live in a hundred sources, from registries to PDFs to supplements to figure panels, and none of the sources describe their findings the same way. While industries like travel planning have structured data and flight or hotel aggregators, biology lacks both standardization and aggregation.

LLMs should read messy sources, suggest mappings, line up fields that mean the same thing, and flag conflicts for a human to review, but the shared format and hub must come first. As data lands in this structure, simple checks keep things comparable: units and scales match, time frames line up, patient groups are matched, and endpoints mean the same thing. Every fact keeps its source, date, and version, so teams can trace provenance and monitor changes. And should be built in a format that is graph-native so platforms can follow and justify links and support the networked reasoning biology requires.

Build models that reason over networks: Networks underpin biology, so our systems must reason over links, not just lines of text. Defensible answers require walking through several connected steps, so-called “multi-hop reasoning”. Modern LLMs are strong at single lookups and fluent summaries, but when multi-hop reasoning is required to respond to a user, they often lose track of constraints, mix cohorts, or skip checks.

At the system layer, perhaps that means moving from “chain-of-thought” to “web-of-thought”: approaches that can traverse multiple paths in a graph at once, check constraints, and reconcile conflicts before proposing an answer.

At the model layer, we might give the model a map of these connections and make those links matter. Instead of treating everything as words in a row, the model should score and select hops (e.g., from a target to its pathways, then to assays and cohorts) and weigh how strong each link is. Along the way, we can encode simple and transparent requirements, so relationships shape the result, not just nearby text. Adopting these principles should enable us to get answers that are traceable, comparable, and reproducible, the qualities high-stakes biotech decisions require.

Figure 3: Chain-of-thought vs. Web-of-thought: Left: a linear “chain-of-thought” compresses evidence into one sentence, ignoring dependencies and hidden constraints. Right: a web-of-thought maps multi-hop links and conditions for action. Biology is a network, not a list, so credible answers require multi-hop reasoning with checks and citations, not a single pass-through text.

Create a user experience that provides validation and feedback loops: Unlike coding LLMs, where feedback is immediate, in biotech the impact of a single choice may not be clear until years later in the clinic or market. Reinforcement learning solved similar challenges through simulation, as in AlphaGo and AlphaZero, but full simulations of biotech investment decisions are not feasible. Adversarial AI, such as digital twin investors that stress-test recommendations, is a possible long-term path but remains speculative. A more immediate approach is to use interaction patterns as training signals. Analysts already perform checks and cross-validations that could be standardized. Systems can learn from explicit corrections as well as implicit behaviors: which trial comparisons are explored or ignored, which analyses are saved or discarded, and how experts navigate between drug mechanisms. In this way, the interface itself becomes validation infrastructure, creating feedback loops that biotech has lacked, offering a new path to make AI trustworthy for guiding high-stakes decisions. Taken together, these principles shift biotech analysis from scattered, manual workflows to something closer to the “single-click diligence” vision. Modernized data pipelines provide standardized, queryable inputs. Network-native models generate multi-hop insights that reflect the interconnected structure of biology, and validation built into the user experience ensures systems continuously refine their domain expertise. The result is not just fluent AI, but trusted decision infrastructure capable of analyzing and supporting the high-stakes choices that define biotech investing and strategic decision making. 

Why we built LABI

As scientists, operators, and investors, we make decisions that demand comprehensive coverage of the data landscape. We need foundational biology, clinical results, regulatory shifts, and more, distilled into clear signals and continuously updated to reflect the dynamic nature of the data. Over the past two years, we have tested existing analytical and AI tools in our workflows and observed firsthand what many across the industry also recognize: generic systems fall short in biotech. 

This gap is why we built LABI, Luma Group’s AI for Biotech Intelligence. LABI is an end-to-end platform purpose-built for biotech, designed to deliver comprehensive coverage of all critical data sources, translate that information into validated, traceable, decision-ready insights, and ensure those insights are maintained in real-time as new evidence becomes available.

Our team has been addressing the foundational limitations of applying generic AI to biology by building a biology-native platform. Rather than augmenting existing workflows with off-the-shelf tools, we are designing LABI from the ground up to capture the full complexity of the field.

LABI aggregates, curates, and harmonizes critical data pools, including peer-reviewed manuscripts, patient-level datasets, biorepositories, clinical trial records, regulatory filings, and commercial and financial datasets, spanning millions of sources. This enables LABI to speak the language of biology (graphs, tables, statistical analyses, and more) and to reason in a highly networked, real-time manner, so insights are both comprehensive and accurate.

At its core, LABI makes biology navigable. It organizes genes, proteins, pathways, samples, and outcomes into a connected map, so comparisons are valid. Every answer is anchored in evidence and linked directly back to sources to eliminate the frustration and dangers of hallucinated outputs. Agentic AI workflows continuously scan and cross-check new information, surfacing what has changed, when it changed, and why it matters. The platform incorporates feedback loops so that every interaction strengthens evidence prioritization, sharpens comparisons, and hardens checks, with improvements compounding over time.

 LABI is the platform we long wished we had as biotech investors, a platform that helps uncover the most meaningful data and insights in a field defined by complexity and constant change. Making the wrong choice or missing the right one has enormous consequences: missed signals can cost billions, delay decade-long timelines, and ultimately slow the delivery of life-saving medicines. The impact of this challenge extends far beyond capital allocation, touching everyone in healthcare who makes high-stakes strategic choices, from clinical development to corporate strategy. This is a mission we are deeply committed to, and we look forward to sharing more as it evolves.

Acknowledgements

Luma Group would like to acknowledge Manav Kumar for his thoughtful discussion and contributions to this white paper.

Jamie Kasuboski, Partner at Luma Group

July 2025

There are many perceptions of what a venture capitalist (VC) is. In the biotech space, VCs are enigmatic because many of us never imagined becoming VCs, and many future VCs might have no idea they’re headed down that path either.

My story, like many in biotech, started with a passion to help sick people. I was fortunate to discover my passion at the age of six, when I told my parents that I wanted to be a genetic engineer. I didn’t fully understand what that entailed, but the film Jurassic Park sparked my curiosity. I was fascinated by the idea that nature had invented biological Legos called DNA that could be assembled to create humans, sea slugs, bananas, bacteria, mold, and, most intriguingly, entirely new life forms.

Less than two decades later, I received my PhD in Molecular and Cellular Biology. I loved deciphering nature’s clues and genetic codes to figure out the “how” and “why” in nature’s playbook. However, I didn’t yet know how to translate this knowledge from the lab into life-saving therapies. I continued my research journey post-PhD and eventually landed an industry postdoc position at Pfizer. At Pfizer, I soaked up every fact, lesson, piece of jargon, and process required to take an initial discovery and turn it into a drug. At this moment, something clicked for me: my true passion wasn’t just making discoveries; it was figuring out how to transform those discoveries into medicines capable of helping patients. My time at Pfizer also made it clear that I wasn’t ready to join an organization as large as Pfizer long-term. Thankfully, after years of trying different career trajectories and with the help of some great mentors, it became clear that biotech venture capital uniquely aligned both my personal goals: contributing to life-changing therapies, and professional goals: being able to earn a living while pursuing my personal passion.

As they say, “Do what you love, and you’ll never work a day in your life.”

Introduction: Venture Capital as the Lifeline of Biotechnology

Translating laboratory discoveries into lifesaving treatments for patients is lengthy, requires massive amounts of capital, and is an extraordinarily challenging journey that is mired in constant failures and setbacks. It requires collaboration across multiple stakeholders, from academic researchers, biotechnology companies, and contract research organizations to regulators and pharmaceutical companies. Unlike other sectors, such as technology or manufacturing, biotech innovation rarely progresses within a single organization. Instead, technologies pass through specialized ecosystems and subsectors, undergoing numerous collaborations and handoffs along the way. Additionally, unlike the tech and manufacturing sectors, which scale by repeatedly iterating the same products, biotech and pharmaceutical companies must continuously innovate or acquire innovation due to IP lifecycle constraints. With all these challenges and transitions, venture capital has emerged as a critical niche for shaping and driving this complex process by uniquely tolerating the industry’s substantial financial demands, prolonged timelines, inherent uncertainties, and high failure rates.  

Investing in biotechnology is not for the faint-hearted. The investment characteristics are poorly matched to traditional financial institutions like banks, later-stage private equity, and public equity markets, which favor shorter investment horizons and lower-risk ventures. Without substantial funding on the order of hundreds of millions of dollars, these discoveries and innovations will never translate into life-saving treatments.  Venture capital firms provide the lion’s share of this critical funding, uniquely equipped with the expertise, alignment, capital, and passion, defined here as the endurance to persist over time, not just enthusiasm, to bridge the risky gap between discovery and clinical development.

A Brief History of Drug Development and the Relationship with Biotech VC

Drug development historically involved large pharmaceutical companies undertaking extensive internal research and development, characterized by decades-long timelines, massive investment requirements, and high failure rates. In the 1970s and 1980s, the emergence of biotechnology, particularly the groundbreaking ability to genetically engineer proteins, revolutionized drug discovery. Biotech startups leveraged academic science and nimble innovation to drive new therapeutic breakthroughs, shifting the paradigm from traditional pharma-dominated R&D to a more collaborative and innovative ecosystem. The rapid rise and success of these startups hinged significantly on venture capital contributions, which provided essential early-stage, patient capital to overcome high-risk barriers and bridge the gap between academic research and commercialization.

Together, VCs, researchers, and entrepreneurs brought transformative innovations, such as monoclonal antibodies, gene therapies, RNA-based technologies, and others to market. This laid the groundwork for modern pharmaceuticals and reshaped the slow-moving, conservative market into the innovation-driven, dynamic ecosystem it is today, saving billions of lives in the process.

Figure 1: Notable Examples of VC-Driven Biotechs.

The Innovation Funding Gap in Biotech

Typically, it takes over a decade and billions of dollars to guide a promising discovery from initial research through clinical trials and ultimately into the market.^1,2 Complicating matters further is the phenomenon known as “Eroom’s Law” (Figure 2), highlighting the troubling paradox that, despite technological advances, drug discovery productivity has steadily declined, resulting in escalating costs and diminishing returns.^3,4 Over recent decades, the average cost of successfully bringing a new drug to market has surged dramatically, surpassing $2 billion per approved drug (depending on who you ask).^5,6

These interrelated complexities have widened the innovation and funding gap: while groundbreaking academic research continues to flourish, translating these discoveries into viable commercial products has become increasingly expensive. This rising cost has significantly squeezed traditional pharmaceutical companies, resulting in an intensification of their dependency on external innovation and reducing internal R&D spend. Over the past two decades pharmaceutical companies have started to play an essential role in the clinical and commercial success of emerging therapies, serving as licensors that guide innovative technologies through the final stages of development and onto the market.

Biotechnology companies, whether private or public, often lack the substantial capital, specialized expertise for late-stage clinical trials, and the robust infrastructure required for successful commercialization. As a result, biotech firms commonly out-license or sell their assets to pharmaceutical companies, leveraging pharma’s financial resources and established commercial pathways to bring new therapies to patients’ bedsides. Influenced by Eroom’s Law and economic pressures, pharma companies have grown more conservative, preferring to wait longer before adopting new technologies, effectively becoming gatekeepers for the commercialization of potentially life-saving innovations.

Figure 2: Eroom’s Law depicted by illustrative historical regression of significant increase of drug discovery and development.

Source: Benchling.

It is precisely at this juncture of financial and scientific uncertainty that biotech venture capital plays an indispensable role. Passionate biotech venture capitalists actively embrace these early-stage challenges, investing precisely when others shy away to provide the bridge between groundbreaking scientific research and commercial viability. The importance of this funding model cannot be overstated; venture capital-backed companies have consistently delivered transformative therapies that shape modern medicine, and without them and their portfolio companies, the industry would quickly decline to a fraction of its size.

Venture capital support is not merely financial; it’s a strategic enabler of groundbreaking medical advances that significantly enhance human health outcomes. Through careful, informed, and visionary investments, biotech venture capital fuels innovation pipelines, enables scientific risks, and ensures that transformative ideas are not trapped in laboratories but rather brought effectively to the patients who urgently need them.

Pulse on the Current Market: When Enthusiasm Outpaces Passion

The rapid exit of generalist funds left hundreds of biotech companies facing financial strain, with low cash reserves and short runways, forcing many to scale back or shut down. This large capital void either needs to be filled by investors, which is unlikely given the sheer scale of capital needed, or there will be another market correction, which will entail more shutdowns, trade sales, and M&A opportunities. The mismatch between high burn rates and the long timelines needed for meaningful value inflection is pushing many promising companies into survival mode. Despite short-term pains, this correction has created compelling opportunities, as high-quality companies trade at steep discounts, offering attractive entry points for experienced investors with fresh capital and limited exposure to prior overvaluations.

Biotech capital markets are cyclical and highly sensitive to broader economic shifts, as seen during the COVID-19 era when the XBI, the S&P’s biotech index, nearly doubled in under a year, attracting trillions in capital. This influx, driven largely by generalist investors that lacked the deep understanding required for biotech investing, fueled innovation, accelerated drug development, and inflated valuations. But it also tipped the balance away from fundamentals. When immediate returns failed to materialize, many of these investors quickly exited, redirecting capital to areas like technology, triggering an abrupt market correction that disrupted the biotech ecosystem and led to the recent downturn.

Biotech VCs: The Sherpas of Innovation

Biotech venture capitalists provide more than financial resources. They serve as skilled sherpas, guiding startups through the challenging journey from early discovery, through clinical development, and ultimately to commercialization. Like Tenzing Norgay, the legendary Nepalese-Indian sherpa who helped Sir Edmund Hillary summit Mount Everest in 1953, the best VCs help startups navigate difficult terrain, carry critical burdens, and stay oriented toward the summit. They bring not only capital but also specialized experience, expansive networks, and strategic insight to each stage of a company’s evolution. So, pick your investors wisely.

VCs also fulfill an essential role as aligners. Biotech startups move through distinct phases, each with different personnel, structures, and missions. VCs ensure continuity. Moreover, effective VCs act as amplifiers, extending a company’s reach and influence within the broader ecosystem. This involves connecting startups with capital sources, facilitating interactions with pharmaceutical stakeholders, advocating externally to enhance the company’s visibility and reputation, and other critical processes.

Above all, the most distinguishing trait of the best biotech VCs is their deep-rooted passion. Many have years or decades of firsthand experience in research laboratories, biotech startups, or pharmaceutical companies, giving them an enduring resilience and ability to maintain unwavering enthusiasm despite the setbacks endemic to biotech innovation. This passion transcends mere enthusiasm; it embodies the capacity to persist through great adversity, remaining committed to the ambitious goal of bringing groundbreaking science to patients in need.

This mission requires VCs to be hands-on, guiding their portfolio through diverse challenges. With dozens of companies under management, VCs must tailor their guidance to each one. This demands not just expertise but time, attention, and strategic dexterity. As VC firms scale, managing their portfolios becomes more complex. Biotech investments often span decades, meaning a single VC firm may find itself managing multiple funds (often two to five at once), each with its own portfolio, LP base, and strategic objectives. The burden of guiding multiple companies across different stages while maintaining internal coherence stretches firms’ bandwidth and heightens the risk of misalignment.

It can be easy to overlook that venture capitalists operate their own business too. In addition to providing tailored support and resources to numerous portfolio companies, VCs must fundraise, manage investors, and oversee firm-level operations. Balancing these dual roles, shepherding others while running a firm, is an often-overlooked challenge. Additionally, VCs are still groups of people that can be prone to errors in decision making, risk calculation, or several other human shortcomings. The best biotech VCs understand this and try their best to stay aligned to fundamentals and remain disciplined even when others are not.  In short, they try to over-index on passion to help patients versus chasing short-term returns with enthusiasm.

Built for Biotech, Equipped for Impact

The world of biotech investing is as diverse as the companies and people who make up the industry. Historically, the most successful biotech investment firms have not been generalists, but specialists whose passion and focus mirror those of the scientists and entrepreneurs driving innovation. Most of these biotech funds take specialization even further, focusing their skills and strategies on specific niches within the broader biotech ecosystem.

We founded Luma Group with these same guiding principles, tailoring our strategy explicitly for the biotech industry and our portfolio. One of our initial decisions was to align our fund’s capital cycle with biotech product development timelines. Specifically, we launched a 15-year fund rather than the traditional 10-year structure. This choice reflects the reality that biotech development cycles often require more time than technology or other sectors and forcing a 10-year investment horizon onto biotech would be fundamentally mismatched. At the heart of our philosophy is a simple, unwavering belief: if we do our job, fewer patients will suffer tomorrow. Guided by this North Star, we set out to build a firm uniquely positioned to achieve such an ambitious goal.

Given the highly regulated and empirically driven nature of the biotech sector, it should come as no surprise that the greatest predictor of success is rigorous, accurate science. Amid all the uncertainties, getting the science right is predominantly about running the correct experiments informed by historical data. This principle has shaped a key mantra within our group: “Better data leads to better decisions, which leads to better outcomes for patients.”

This mantra is what guided Luma Group to establish a dedicated research division within our funds, supported by dozens of KOLs and advisors, from former pharmaceutical CEOs and top regulatory officials to PhDs and postdoctoral researchers. The breadth and depth offered by this network ensures coverage of every critical vertical in our portfolio and fund exposure:

• Academic Experts provide visibility into groundbreaking innovations and technologies
• Discovery and Development Specialists focus on execution and translating early-stage technologies into next-generation therapeutics, medical devices, or diagnostics
• Clinical and Regulatory Advisors assist in navigating the rapidly evolving clinical and regulatory landscape
• Commercial Experts help us understand pharmaceutical decision-making processes and broader market dynamics affecting our portfolio companies

To complement this network, we have developed a next-generation research platform built around proprietary analytical software known as LABI (Luma AI Brain Initiative). This platform provides both our fund and our portfolio access to trillions of curated data points through an intuitive interface that significantly reduces diligence times while leveraging advanced meta-analysis and analytics, empowering us to uncover insights others typically miss.

With these foundational elements in place, Luma Group has strategically aligned its passion and fund lifecycle with the asset class and companies that are driving innovation. With a lot of passion, hard work, and some luck, Luma Group can drive meaningful improvements in patient outcomes within our lifetime.

1. Pisano, G. P. (2006). Science Business: The Promise, the Reality, and the Future of Biotech. Harvard Business School Press. ↩︎
2. Booth, B. L., & Zemmel, R. W. (2004). Prospects for productivity. Nature Reviews Drug Discovery, 3(5), 451–456. ↩︎
3. Scannell, J. W., Blanckley, A., Boldon, H., & Warrington, B. (2012). Diagnosing the decline in pharmaceutical R&D efficiency. Nature Reviews Drug Discovery, 11(3), 191–200. ↩︎
4. Deloitte Centre for Health Solutions. (2022). Measuring the Return from Pharmaceutical Innovation 2022: Balancing the R&D equation. Deloitte Insights. ↩︎
5. DiMasi, J. A., Grabowski, H. G., & Hansen, R. W. (2016). Innovation in the pharmaceutical industry: New estimates of R&D costs. Journal of Health Economics, 47, 20–33. ↩︎
6. Deloitte Centre for Health Solutions. (2022). Measuring the Return from Pharmaceutical Innovation 2022: Balancing the R&D equation. Deloitte Insights. ↩︎

Jamie Kasuboski, Partner at Luma Group and Rob Plasschaert, Senior Director of Biology at Stealth Newco

Innovation in biotechnology is driven by uncovering novel biological insights and translating them into life-saving therapeutics, diagnostics and medical devices. Over the past two decades, breakthroughs have largely stemmed from analyzing vast biological datasets, such as those generated by human genome projects.

Today, advancements in artificial intelligence (AI) and machine learning (ML) have significantly enhanced our ability to systematically analyze massive datasets, identifying complex relationships across genomic, proteomic, transcriptomic, metabolomic and other data simultaneously. The cross-section of all of these “-omics” is what we define as multi-omics, which represents a large untapped domain for future biotech innovation.

The convergence of affordable, sophisticated AI/ML analytics and large-scale multi-omics data collection has marked a pivotal shift within biotechnology, from single-omics approaches to integrated multi-omics innovations.

June 2025

Introduction: Multi-Omics and the Era of Big Data in Biotechnology

Over 20 years after the first published human genome, biotechnology is now firmly a discipline of big data. The dissection of disease mechanism is done by creating a logical path of cause and effect that moves from origin (e.g. a genetic mutation) to disrupted biology process (e.g. non-functional protein and pathway) to presentation of clinical symptoms (e.g. cancer). Traditionally, the scope of this work has been limited by scale, and progress has been consistent but slow. Advances in methodology have transformed full-coverage “-ome” profiling—genome, transcriptome, epigenome, metabolome, and beyond—from bleeding-edge novelty into standard practice, offering more qualitative quality control. We can now routinely generate terabytes of molecular measurements from a single study. We, and many others, believe that connection between these large datasets will define the next era of multi-omics drug development. This emerging field focuses on the integration and analysis of large-scale molecular and clinical data, enabling the systematic dissection of biological cause-and-effect at scale. By viewing molecular disease through this multi-faceted lens, researchers can identify and translate novel insights into new therapeutics, diagnostics and medical devices.

Looking through the Compound lens of Multi-Omics

By combining large datasets that characterize disease etiology with clinically meaningful endpoints, multi-omic analysis is poised to deliver transformative insights. High-throughput profiling of the central dogma—DNA → RNA → protein—is now routine, and assays for modulators (e.g., epigenetic marks) and downstream effectors (e.g., metabolites) have dramatically decreased in cost while increasing in sensitivity, rendering them nearly run-of-the-mill. Fifteen years ago, whole-genome sequencing cost over $10 million per genome, RNA-seq was low-throughput, proteomics relied on 2D gels, and electronic health records (EHRs) remained siloed; today, sequencing runs under $500 per sample, single-cell multi-omic kits can simultaneously profile chromatin accessibility and gene expression, and modern proteomics platforms quantify thousands of proteins in a single day.

Clinically, routine laboratory tests, digital histopathology, and imaging now feed into AI-enabled pipelines that extract multi-scale features, while EHR data—once trapped behind Epic or Cerner—are routinely exported as de-identified OMOP/FHIR–formatted datasets via research data warehouses. Public resources such as MIMIC-IV, NIH’s All of Us Research Program, and the UK Biobank exemplify how ICU telemetry, standardized lab values, and de-identified clinical notes can be linked to genomics and metabolomics under strict governance. What once required bespoke protocols, custom ETL pipelines, and extensive manual annotation has evolved into a streamlined, plug-and-play ecosystem, enabling researchers to integrate multi-omic and clinical data seamlessly and uncover biological insights that were impossible to detect a decade ago. All of these sources provide a rich array of data spanning numerous dimensions, from molecular-level insights to comprehensive patient health journeys.

Table 1. Omic Modalities and Their Captured Insights

Category	Data Modality	What it captures and/or quantifies
Molecular Mechanisms	Genomics	The genetic blueprint of an organism’s genome
	Epigenomics	Reversible chemical marks (e.g., DNA methylation, histone modifications) that modulate gene expression
	Transcriptomics	Dynamic gene-expression programs (mRNA abundance and isoforms)
	Proteomics	Proteins, their splice variants, and post-translational modifications
	Metabolomics	Small-molecule metabolites whose levels change with cellular activity and stress
Disease & Clinical Outcomes	Radiology & Functional Imaging	MRI, CT, PET, ultrasound imaging that quantify disease states over time in various organs
	Digital Pathology & Spatial Slides	Whole-slide histology, multiplex immunofluorescence, spatial transcriptomics mapping cellular phenotypes to anatomical context
	Electronic Health Records (EHR)	Structured labs, vitals, medications, procedures, plus unstructured clinical notes collected across years of care
	Longitudinal Laboratory Panels	Serial hematology, chemistry, and biomarker tests (e.g., HbA1c, troponin) tracking disease progression or therapeutic response

Why is Multi-Omics Poised to Make an Impact Now?

Multi-omics approaches use multidimensional datasets whose complexity often surpasses the capabilities of classical statistical methods. Although earlier computational approaches were effective, recent advances in AI/ML have not only dramatically increased computational power but also enable seamless integration across multiple datasets—each with its own unique architecture. By integrating neural networks—particularly advanced deep‐learning architectures, graph neural networks, and probabilistic causal frameworks—researchers can now uncover insights and identify connections that were too subtle for earlier computational methods, turning analytical challenges into strengths and offering a powerful means to decipher biological complexity. AI models integrate heterogeneous data modalities—such as DNA variants, RNA expression counts, protein abundances and metabolite concentrations—into unified latent representations, preserving critical biological interactions across layers. For example, alignment models facilitate a comprehensive understanding of complex systems by embedding different data types into shared latent spaces, maintaining biological coherence across diverse “-omic” layers.¹ Researchers have begun to apply these techniques to profile immune cells directly from patient samples. For example, Dominguez Conde et al. (2022) took early steps toward characterizing immune cells in both healthy individuals and diseased patients, aiming to understand how their multi-omic profiles shape the immune system’s adaptation and function in different tissue environments.²

Moreover, AI-based methods significantly improve data quality through noise reduction and imputation. Techniques, such as autoencoders and diffusion models reconstruct missing values, correct batch effects and enhance the signal-to-noise ratio in noisy assays. Variational autoencoders, for instance, have been successfully employed to impute missing data in single-cell multi-omics, dramatically enhancing analytical robustness.³ Additionally, supervised deep-learning models trained on clinical endpoints—including patient survival, relapse rates and therapeutic response—can accurately link complex molecular patterns to clinically relevant outcomes. These models distill intricate biological signatures into actionable insights, thereby accelerating precision medicine initiatives and facilitating personalized therapies (Lee et al., 2020).⁴

Foundation models trained on extensive multi-omics are increasingly valuable for generating testable biological hypotheses. These models predict causal interactions, protein structures and even simulate the effects of targeted genetic or pharmacological interventions. For instance, AlphaFold and similar AI systems demonstrate how computational predictions can effectively precede laboratory validation, dramatically shortening the cycle from data collection to biological discovery.⁵ AI is a critical translator, converting dense, molecular-level information into meaningful clinical insights and actionable therapeutic strategies, thereby bridging the gap between complex multi-omics data and tangible patient benefits.

Table 2. Standard Omic workflow

---

A Case Study in Low-Throughput Translation: γ-Secretase in AD

Consider γ-secretase inhibition in Alzheimer’s disease. Decades of biochemistry showed that γ-secretase generates amyloid-β peptides that aggregate into neurotoxic plaques; blocking the enzyme seemed like a slam-dunk therapeutic strategy. Yet Lilly’s semagacestat—a potent γ-secretase inhibitor—failed spectacularly in Phase III trials. Cognition worsened faster than in placebo, and adverse events spiked. One idea is that Amyloid processing is only one facet of a vast neurodegenerative network; γ-secretase also cleaves Notch receptors and other crucial substrates. If researchers had a multi-omic, systems-level view of neuronal biology—linking genomic risk alleles, transcriptomic stress responses, proteomic pathway crosstalk and metabolic dysfunction—they might have predicted these liabilities before thousands of patients were exposed. The lesson is clear: single-node interventions based on incomplete models can backfire.

Challenges That Remain

Although early precision medicine successes emerged from single-omic approaches, multi-omics strategies—despite their promise—face critical implementation hurdles. Foremost is data quality and standardization: unlike genomic sequencing’s unified formats, multi-omics suffers from inconsistent sample collection, processing methods and metadata curation, limiting cross-study comparability. A further obstacle is interpretability; predictive models often function as “black boxes,” failing to provide the transparent mechanistic insights regulators and clinicians require for trust and actionable decisions. Lastly, the scalability of experimental validation remains constrained, with wet-lab confirmation via phenotypic screening, organoid systems and CRISPR perturbations lagging far behind computationally generated hypotheses.

These bottlenecks—poor data standardization, opaque models and limited validation—represent significant barriers to realizing multi-omics’ clinical potential. Even with these challenges, research continues to make progress on removing these bottlenecks and with new predictive models combined with more efficient and standardized wet lab data collection. Additionally, new AI/ML approaches help fill in the bottlenecks and will be explored in a later section.

Table 3. Current Omic Bottlenecks and Pain Points

2. AI & Machine Learning make Multi-Omics Possible

AI tools are rapidly reshaping the landscape of biomedical research by solving longstanding challenges in data analysis, interpretation and utilization. Traditional methods in biology and medicine frequently face bottlenecks related to scale, accuracy and speed, limiting discovery and clinical translation. Today’s AI-powered tools offer unprecedented precision, automation and analytical depth, poised to resolve critical choke points throughout the multi-omics workflow—from initial raw signal cleanup to sophisticated drug design. Below, we highlight specific examples of the impact these AI tools are already making, along with a selection of emerging use cases. (For a deeper dive into our philosophy—more data isn’t better data; curated data is better data—see our prior AI white paper.)

For instance, one of AI’s transformative capabilities lies in converting biological sequences into accurate three-dimensional protein structures. Traditionally, researchers relied on experimental techniques like wet-lab protein crystallography, a process that could take months per protein and left most proteins structurally unresolved. AI-based solutions, exemplified by AlphaFold, have dramatically changed this reality. AlphaFold has generated readily accessible structural models for nearly 200 million proteins, empowering vaccine developers and enzyme engineers to rapidly obtain atomic-level detail in seconds instead of months.⁶

AI also significantly improves the quality and usability of genomic data. While advanced sequencing technologies such as long-read sequencers deliver critical insights, they often come with higher error rates compared to short-read counterparts. AI-driven models, including Google’s DeepVariant, address this issue by effectively “cleaning” raw genomic reads and boosting variant-calling accuracy to near-clinical standards.⁷ Such tools dramatically reduce the manual quality control burden and time—shaving weeks off the analysis pipeline and enabling faster translation from genomic discovery to clinical action. Additionally, AI facilitates the annotation and interpretation of complex single-cell datasets, a process that is notoriously labor-intensive and prone to subjectivity. Traditional manual annotation of million-cell datasets is both slow and variable across annotators. AI-driven solutions, such as the open-source popV ensemble, systematically assign cell-type annotations along with confidence scores. This automated process highlights only the ambiguous 10–15% of cells for expert review, significantly accelerating workflows and ensuring higher consistency and reproducibility across analyses.⁸

AI excels at integrating multi-omic data streams—such as DNA, RNA, and digital pathology—to create comprehensive predictive models. While individual biomarkers often fail to capture the complexity of diseases, AI-based multi-omic fusion models achieve remarkable accuracy. Recent pan-cancer studies employing deep-learning techniques have successfully combined diverse data types into unified survival risk scores. These AI-derived scores have consistently outperformed traditional stage-based predictions, delivering superior prognostic accuracy across studies involving more than 15,000 patients.⁹

Collectively, these advancements demonstrate AI’s potential to transform biomedical research, delivering faster, more precise and clinically relevant insights at unprecedented scales.

Table 4: Real-World Challenges and Examples of AI-enablement

Pain Point	What AI Can Do	Everyday Example
Too much data, not enough insight	Spots hidden warning-sign patterns that clinicians would never have time to sift out manually.	A machine-learning screen of newborn blood samples uncovered a 4-gene “early-warning” fingerprint for sepsis—flagging babies days before symptoms appeared.10
Different hospitals use incompatible equipment	Lines up images or lab results from many sites so they can be compared as if they came from one scanner or one lab.	An AI harmonization tool lets researchers combine breast-MRI databases into one study, boosting the accuracy of tumor-detection software across both hospitals.11
Important signals are buried in noise	Cleans and sharpens data, filtering out scanner glitches or stray measurements.	In lung-cancer screening, an AI system that de-noised CT scans spotted malignant nodules earlier and with fewer false alarms than expert radiologists.12
Key test results are missing	Predicts likely values or tells clinicians which single test would add the most value, cutting down on repeat blood draws.	A study showed that AI imputation could reliably fill in missing lab results in electronic health-records for stroke and heart-failure patients, improving risk models without extra testing.13
Translate big data into clinical outcomes difficult	Converts continuous streams from wearables into medically meaningful alerts.	The 400,000-participant Apple Heart Study used an AI algorithm in a smartwatch to flag atrial-fibrillation episodes with 84 % accuracy, prompting users to seek timely care.14

3. Factors Driving Growth in Multi-Omics & AI

Three forces—capital, capability and clinical pull-through—are reinforcing one another and accelerating adoption of multi-omics platforms and solutions in today’s market environment.

I) Table 5: Plentiful capital for big data bioscience

What’s Happening		Why it Matters	Example
Capital for AI platforms is abundant	Investors see multi-omics + AI as the next big moment for biotech		Global multi-omics platform revenue is expected to nearly double—from ≈ $2.7bn (2024) to $5bn (2029)15
Big rounds for data-centric start-ups	Larger war-chests let companies build both wet-lab and compute infrastructure		The 20 biggest biotech start-ups raised $2.9bn in Q1 2024, many with AI/multi-omics pitches16
Generative-AI boom spills into life-sciences	General-purpose GenAI tools lower barrier to sophisticated modeling		Venture funding for GenAI hit $45bn in 2024, up ~2× YoY17
Strategic pharma partnerships	Pharma licenses data access and co-develops AI platforms instead of building in-house		UK Biobank’s new AI-ready proteomics program launched with 14 pharma partners

II) Capabilities for data generation and analysis continue to improve

The cost and economics of generating multi-omics data are changing rapidly: the price of whole-genome sequencing, once counted in the thousands of dollars, is now approaching the USD 200 threshold promised by the latest high-throughput instruments, effectively removing cost as the principal barrier to large-scale genomic studies.¹⁸ Parallel progress in proteomics has reduced mid-plex panel costs to well under USD 20 per sample, broadening access to routine protein profiling. At the same time, data resources have expanded in both depth and breadth. The UK Biobank has released metabolomic measurements for approximately 121,000 participants—and complementary panels quantifying roughly 3,000 plasma proteins—thereby creating an unprecedented reference for population-scale multi-omics analyses.^19,20 These volumes would be unmanageable without a concurrent maturation of cloud-computing infrastructure. On-demand GPUs and browser-based “auto-ML” notebooks now allow investigators to execute multi-omic workflows that once required institutional high-performance clusters, placing advanced analytics within reach of modestly resourced laboratories. Finally, the regulatory climate is becoming markedly more receptive. Recent FDA guidance on the use of real-world evidence and tissue-agnostic companion diagnostics explicitly acknowledges integrated molecular signatures as acceptable decision-making inputs, thereby creating a clearer path from multi-omic discovery to clinical implementation.

III) Table 6: Early efforts point to big possible successes in the space

4. Luma Group’s Position and Vision

Scientific progress hinges on more than just data—it depends on the ability to make sense of it to improve outcomes for patients. At Luma Group, we invest in companies that are redefining how data is used to shape the future of drug discovery and development. While modern research tools can now produce unprecedented volumes of biological information—from single-cell sequencing to proteomic and metabolomic profiling—sheer quantity doesn’t guarantee clarity. The true advantage lies in the ability to connect disparate data streams, uncover hidden patterns and translate them into actionable insights.

For complex diseases, a holistic understanding of interrelated datasets is crucial to deciphering disease biology. Ultimately, these data function as an interconnected system, and forward-thinking companies increasingly rely on AI and ML to analyze these enormous datasets, revealing the complexities of many unmet medical needs and opening the door to new breakthroughs. We believe we are transitioning out of the genomics era and into the multi-omics era—one where integrated datasets and advanced analytical tools will transform the way we discover, develop and deliver the next generation of therapeutics.

Luma Group continues to champion this new era of innovation by focusing on companies that harness large multi-omics datasets alongside advanced AI/ML approaches. One of our earliest investments, Rome Therapeutics, built its discovery engine around an often overlooked repeatome, a human genome section that does not code for a human protein. By mining large patient datasets, the Rome team pinpointed a key correlation from the repeatome and pathways involved in autoimmunity, ultimately uncovering LINE1 as a novel target with broad therapeutic potential across multiple autoimmune indications.

We invested in Curve Bio because they applied similar multi-omics analysis to diagnostic applications. Their AI/ML-powered platform sifts through massive datasets to detect subtle changes in the methylation patterns of cell-free DNA—changes that strongly correlate with disease progression, particularly in liver tissues. These insights exceed standard-of-care detection and other methods in both sensitivity and selectivity, offering significant promise for earlier and more accurate diagnoses.

Our investment in Character Biosciences represents an archetype of Luma’s multi-omics investment strategy. As we recently detailed in a separate piece, Character Bio integrates ocular imaging, genetic profiles, metabolomics and patient natural histories to study dry AMD. The company uncovered novel biological pathways and therapeutic targets by applying AI/ML techniques to these massive datasets. With two lead assets set to enter the clinic within the next 12 months, Character Bio is on course to become one of the first to deliver approved therapies built on a multi-omics foundation, powered by advanced AI/ML analysis.

Our fund is optimistic that innovation within our sector will continue to grow as we leverage large multi-omics datasets with advanced AI/ML. At the heart of this growth are new, innovative tools and approaches that will lower the cost of generating these extensive datasets beyond just single-omics. We have begun to see a shift in how large consortiums are expanding their “-omics” footprint beyond just genomics. One prominent initiative that has embraced the power of multi-omics is the UK Biobank. This program has enrolled over 500,000 volunteers who will donate information—including biological samples, physical measurements, body and brain imaging data, bone density data, activity tracking and lifestyle questionnaire data—over the span of 30 years. Beyond genomics, the Biobank collects proteomic, metabolic, MRI imaging, natural history and other key datasets to holistically understand how these historically disparate data types interact. Their goal is to translate these insights into novel findings that can inform the development of new therapeutics and diagnostics.

Over the last decade, we have seen other private and public initiatives, such as the All of Us initiative and Project Baseline, set out to gather similarly large multi-metric data for the same purpose. The key to capitalizing on these datasets lies in understanding subtle and often hidden connections within them—an approach made possible through AI/ML methods that can identify insights too complex or minute for human intuition alone. In our portfolio companies, we have observed how crucial AI/ML approaches are for extracting valuable information and maximizing the potential of these datasets. This trend—integrating AI/ML with the collection of massive datasets through new and innovative tools—will likely continue and, in doing so, provide patients with a new generation of therapeutics and diagnostics.

5. Conclusion and Outlook

We have witnessed how impactful single-omics analyses, particularly genomics, have been in understanding disease pathology and progression, leading to dozens of approved drugs. However, single-dimensional “-omic” data inherently has limitations given the complexity of diseases and disorders we aim to treat. The initial wave of omics-based medicine was primarily driven by advances in genomic sequencing technologies and substantial reductions in sequencing costs, fueling significant innovation over the past two decades. Now, we see a similar technological advancement unfolding in other “-omics” domains—including proteomics, metabolomics, glycomics and beyond—with costs starting to decline in a manner comparable to genomics, setting the stage for further innovation. Yet, aggregating, managing and analyzing massive multi-omics datasets consisting of billions of data points poses unique challenges, necessitating more sophisticated artificial intelligence and machine learning methods. Both private and public sector initiatives are already emerging to address these challenges.

In the coming decade, we anticipate a new wave of innovation in multi-omics medicine, potentially surpassing the transformative impact initially driven by genomics. Luma Group aims to actively invest in and nurture this exciting frontier, positioning itself at the cusp of this transformative era in multi-omics medicine. If you’re building in the space, please reach out to us.

---

1. Argelaguet, R., et al. (2021). Multi-omics integration approaches for disease modeling. Nature Reviews Genetics, 22(6), 345–362. ↩︎
2. https://www.science.org/doi/abs/10.1126/science.abl5197 ↩︎
3. Lotfollahi, M., et al. (2022). Mapping single-cell data to reference atlases by transfer learning. Nature Biotechnology, 40(1), 121–130. ↩︎
4. Jumper, J., et al. (2021). Highly accurate protein structure prediction with AlphaFold. Nature, 596(7873), 583–589. ↩︎
5. Jumper, J., et al. (2021). Highly accurate protein structure prediction with AlphaFold. Nature, 596(7873), 583–589. ↩︎
6. https://www.nature.com/articles/d41586-022-02083-2 ↩︎
7. https://pmc.ncbi.nlm.nih.gov/articles/PMC11466455/ ↩︎
8. https://www.nature.com/articles/s41588-024-01993-3 ↩︎
9. https://www.nature.com/articles/s43018-024-00891-1 ↩︎
10. https://www.niaid.nih.gov/news-events/gene-signature-at-birth-predicts-neonatal-sepsis-before-signs-appear ↩︎
11. https://pmc.ncbi.nlm.nih.gov/articles/PMC8508003/ ↩︎
12. https://news.northwestern.edu/stories/2019/05/artificial-intelligence-system-spots-lung-cancer-before-radiologists/ ↩︎
13. https://www.nature.com/articles/s41746-021-00518-0 ↩︎
14. https://www.nejm.org/doi/full/10.1056/NEJMoa1901183 ↩︎
15. https://www.bccresearch.com/market-research/biotechnology/multiomics-market.html?srsltid=AfmBOordStMDYvNp3BPq_s_wZgT3nGfnzBzGpzJrUp4Wd1VNObRVqVz1 ↩︎
16. https://www.drugdiscoverytrends.com/20-biotech-startups-attracted-almost-3b-in-q1-2024-funding/ ↩︎
17. https://www.mintz.com/insights-center/viewpoints/2166/2025-03-10-state-funding-market-ai-companies-2024-2025-outlook ↩︎
18. https://www.biostate.ai/blog/genome-sequencing-cost-future-predictions ↩︎
19. https://www.nature.com/articles/s41597-023-01949-y ↩︎
20. https://www.ukbiobank.ac.uk/enable-your-research/about-our-data/past-data-releases ↩︎

The convergence of various factors, including advancements in OMIC data, repurposing of precision oncology technologies and high demand from patients and pharmaceutical companies, has positioned inflammation and immunology sectors for explosive growth, driving the development of safer and more effective precision therapies.

March 2024

Inflammation and Immunology Disease Landscape

Inflammatory and Immunology (I&I) diseases affect approximately 35 million Americans or over 10% of the population. Despite the prevalence, current treatments are ineffective in managing these conditions.

Unsurprisingly, this lack of effective treatments has led to significant demand for new and innovative therapies that are safe and efficient. However, biotech and pharma R&D focus to date has largely been on more severe indications (e.g., oncology), which despite a smaller size, have seen rapid market growth. Precision therapy methods, harnessing established and de-risked technologies derived from precision oncology, hold the promise of catalyzing a comparable revolution in the treatment of I&I diseases. This presents an opportunity for enhanced disease management and ultimately, superior patient outcomes.

The lack of innovation in I&I drug development has been a critical factor contributing to the slower growth of the I&I market. Interestingly, despite the two times larger patient population than oncology, I&I diseases only generate 60% of the market size (Figure 1).

Figure 1: Comparison of I&I disease and oncology markets.

Source: SEER, NIH, IQVIA, Global Data, Evaluate Pharma.

This discrepancy cannot be attributed to differences in treatable disease subtypes or drug treatment costs, but instead to the limited treatment options available for I&I diseases. Limited options have resulted in redundant use cases and lower overall market penetration.

In comparison, oncology has more than twice the number of marketed drugs, which are more diverse in their mechanisms of action and use cases. For example, the top 16 selling oncology drugs generated over $50 billion in sales in 2022 alone according to Global Data, accounting for approximately 50% of the total sales, and 13 out of 16 of these therapies have novel biological mechanisms of action. On the other hand, the I&I market is dominated by only five drugs that comprise around 50% of the market according to Global Data and have only three unique biological mechanisms of action. This dynamic highlights the need for innovation and the potential for market expansion in the space.

The historical safety issues in I&I drug development have been a primary reason for the lag in the market’s growth. Many promising drugs that have demonstrated efficacy in preclinical and clinical trials have failed to enter the market due to safety concerns associated with chronic use. A prime example is CD20 targeting antibodies like Rituxan (rituximab) have shown promising efficacy in the clinic for multiple I&I indications, the class of drugs harbors severe safety issues/risks (opportunistic infection, anaphylaxis, acute coronary syndrome and even death) with chronic use. This “safety Achilles heel” results from target redundancy in disease and healthy cells/tissues, where many tractable drug targets overlap with those expressed in healthy states.

The need for more innovation in addressing these safety concerns has hampered the progress of I&I drug development, with only Stelara (an anti IL-12/IL23 antibodies used for the treatment of Crohn’s disease, ulcerative clotisese, plaque psoriasis and psoriatic arthritis) out of the top five selling I&I drugs entering the market in the past 15 years. This lack of innovation has significantly impacted the growth and potential of the I&I market. However, it highlights a potential opportunity to drive market growth with novel approaches.

Looking forward, I&I drug development can learn from the similar challenges experienced in oncology drug development in the 1990s and early 2000s, which were overcome by shifting to a personalized medicine approach.

Chemotherapy drugs (i.e., antineoplastics) were the legacy standard of care in oncology and have broad mechanisms of action, systemically killing rapidly dividing cells. While this can successfully target cancer, these cells also exist in several non-cancerous tissue types like hair and the lining of the stomach, resulting in the severe side effect profile that is ubiquitous in the class. The non-specific targeting of cells created an immense treatment burden on patients and spurred the improvements in specificity, efficacy and safety seen in the last decade. New scientific insights from collecting and analyzing massive patient datasets enabled this shift, paired with innovative technologies that allowed therapies to target tumor cells while selectively sparing healthy cells. This shift to precision approaches ushered in the era of precision oncology, unlocking enormous market growth that quadrupled in less than 15 years (Figure 2) and vastly improved patient outcomes (Figure 3). I&I drug development can leverage similar strategies to address safety concerns and drive innovation, paving the way for its growth and success in the market.

Figure 2: U.S. oncology sales have increased dramatically over the past 10 years with approvals of immunotherapies (e.g., Keytruda) and other targeted agents (e.g., Tagrisso) driving new sales volume.

Source: Global Data, Evaluate Pharma.

Figure 3: The growth in the U.S. oncology market was enabled by the dramatic improvements in patient care outcomes supporting their clinical success. This is particularly evident in front line EGFR NSCLC.

Note: Front line standard of care prior to TKI approvals assumed to be platinum-based chemotherapy. PFS: Progression-free survival, ORR: Overall response rate. Source: FDA label, SEER, clinical trials.

To overcome the innovation bottleneck and thrive in the market, I&I drug development can take cues from the success of oncology by making two fundamental shifts:

1. Embracing OMICs Large Data Sets: One crucial missing component in developing I&I precision therapies has been the need for novel biological insights from analyzing massive patient datasets. Until now, researchers have lacked the proper tools, such as single-cell sequencing, advanced fluorescence-activated cell sorting and advanced mass spectrometry, to accurately decipher the immune system’s complexities. By leveraging these technologies and analyzing extensive patient data, I&I drug developers can gain valuable insights into disease mechanisms and identify new targets for precision therapies.

2. Repurposing Precision Oncology Technologies: The technologies that revolutionized precision oncology, such as heterobifunctional molecules, bispecific antibodies and cell therapy, can be repurposed for developing precision I&I drugs. In oncology, these innovative technologies have already shown success in targeting tumor cells while sparing healthy cells, and they can also be applied to I&I diseases. By repurposing these proven technologies, I&I drug developers can accelerate the development of safe and effective therapies, overcoming the stagnation in the I&I market.

Investors and companies who focus on analyzing patient data and repurposing precision oncology technologies early on will likely meet the demand for new therapies during the transition to the era of precision therapies in I&I drug development. This success may result in significant therapeutics creation for the large unmet medical need, which are positioned well to gain strong market traction.

I&I Opportunity

Multiple converging factors have positioned the I&I field for rapid growth, potentially surpassing the oncology market. This historically underserved market now has the necessary pieces to overcome historical bottlenecks in I&I drug development and usher in a new era of precision therapies.

One essential advancement is the availability of massive OMIC data, which can be used to address and advance therapeutic development in I&I diseases across several axes. New decoding technologies, like single-cell sequencing, provide novel biological insights for developing precision I&I treatments. This wealth of data has enabled better patient stratification, allowing researchers to characterize patients more accurately and identify responders while excluding non-responders who may pose safety risks. Additionally, this data has revealed new target proteins and cell types that could overcome historical safety issues by increasing drug specificity and expanding the repertoire of potential targets for I&I drug development.

Moreover, I&I drug development can leverage cutting-edge precision oncology technologies that have emerged from the advancements made in the last two decades. These clinically validated technologies, such as adoptive cell therapy and bispecific antibodies, can be repurposed to address safety concerns in I&I drug development. In addition, these modular technologies can be more easily adapted for cost-effective and rapid I&I drug development. For example, companies have repurposed oncology bispecific antibody technology to target immune cells involved in atopic dermatitis while sparing healthy immune cells.

The demand for I&I therapies in the market remains enormous, with many diseases still lacking effective treatments. Many large pharmaceutical and biotechnology players have identified this gap in the market and have been exploring opportunities to access the value. Given their lean and waning I&I drug pipelines, they have looked outwards, lending to several lucrative biotech acquisitions. For instance, companies like Arena Pharmaceutical ($6.7B in 2021), Pandion Therapeutics ($1.85B in 2021) and Momenta Pharmaceuticals ($6.5 billion in 2020) have been acquired by pharmaceutical giants Pfizer, Merck and J&J, respectively, showcasing the appetite for new precision I&I therapies and the potential for continued acquisitions in the future (Figure 4).

Figure 4: Select M&A and licensing I&I deal activity from large cap pharma/biotech since 2020.

Target	Acquirer	Deal Type	Upfront	Total
Televant	Roche	M&A	$7.1 B	$7.1 B
Arena Pharmaceuticals	Pfizer	M&A	$6.7 B	$6.7 B
Momenta Pharmaceuticals	J&J	M&A	$6.5 B	$6.5 B
ChemoCentryx	Amgen	M&A	$3.7 B	$3.7 B
Chinook	Novartis	M&A	$3.5 B	$3.5 B
DICE Therapeutics	Eli Lilly	M&A	$2.4 B	$2.4 B
Pandion Therapeutics	Merck	M&A	$1.85 B	$1.85 B
Garcell Biotechnology	AstraZeneca	M&A	$1 B	$1.2 B
Nimbus Therapeutics	Takeda	License	$4 B	$6 B
EVQQ Therapeutics	Gilead	License	Not disclosed	$658.5 M

Source: Company Filings, Pitchbook.

Overall, the convergence of various factors, including advancements in OMIC data, repurposing of precision oncology technologies and high demand from patients and pharmaceutical companies, has positioned I&I for explosive growth, driving the development of safer and more effective precision therapies.

Luma is strategically positioned as a pioneering investor, ready to capitalize on the burgeoning I&I market. We have extensive expertise and active engagement in this market’s two primary value drivers. Additionally, we have an established presence in the emerging field of precision I&I -focused biotech, with active investments in this rapidly growing sector.

Value Driver #1: Luma’s expertise in identifying essential OMICs technology is central to our investment approach and ingrained in our DNA. Our fund has strategically invested in core technologies fueling the new wave of precision I&I therapies. As a result, we possess a keen understanding of the power of OMICs data sets and how to leverage them for the best drug discovery. This knowledge places Luma in a prime position to identify companies and technologies that are highly differentiated and poised for success.

Value Driver #2: Luma’s proficiency in repurposable precision oncology technologies is supported by a deep understanding and proven track record of funding and founding and investing in multiple precision medicine oncology companies. As a result, we are well-versed in cutting-edge technologies like bivalent small molecules, heterobifunctional molecules, bispecific antibodies and others that can be rapidly repurposed to develop precision I&I therapies (Figure 5). This expertise positions us at the forefront of the intersection of precision oncology and I&I, unlocking new possibilities for innovative treatments.

Figure 5: Precision therapeutics modalities with potential for translation into I&I disease.

Source: Nature, Cell.

Committed to catalyzing the new precision I&I market, our fund is already actively engaged and primed to capitalize on the new wave of I&I disease market growth. For example, in 2023, we invested in ROME Therapeutics – a company pioneering a new class of precision I&I therapies called LINE-1 reverse transcriptase inhibitors (NRTIs) – demonstrating our proactive approach to identifying and seizing first-mover opportunities.

ROME’s groundbreaking approach targets the uncharted territories of the human genome, often referred to as the ‘dark’ or noncoding regions. Utilizing cutting-edge sequencing and analytical methods, they have made remarkable discoveries. Specifically, they’ve identified certain genomic regions intricately linked to the progression of multiple types 1 interferon-driven I&I diseases, such as Lupus, also referred to as interferonopathies.

Their research has led to the identification of LINE-1 reverse transcriptase (RT) as a new therapeutic target. This target operates upstream of already validated targets, offering a unique approach to treatment (Figure 6). Significantly, it minimizes the risk of infection from opportunistic pathogens, given it is not an immunosuppressant. This promising development is rapidly progressing, and ROME is on track to introduce its first drug to clinical trials in 2024. ROME is applying its approach to advance earlier-stage programs across a spectrum of autoimmune diseases as well as neurodegenerative diseases and cancer.

Figure 6: Rome Therapeutics’ mechanism of action versus approved SLE treatments.

Source: Company materials.

Additionally, we recognize the value in and actively seek investments leveraging precision oncology technologies for I&I therapies, further showcasing our commitment to driving innovation and capturing market opportunities.

Luma is well-prepared to lead the I&I investment landscape, leveraging our expertise in OMICs technology and data, repurposable precision oncology technologies and active participation in pioneering I&I companies.

Conclusion

The I&I market is a rapidly expanding investment sector, but a lack of groundbreaking therapies has hindered its growth. Despite this challenge, the foundational elements are now coming together to propel the development of innovative precision therapies and unlock significant market growth. Luma possesses the expertise, experience and enthusiasm to be a pioneering investor in this promising market. Therefore, we are the ideal investment partner to capitalize on emerging opportunities in the precision I&I market, offering the best growth potential.

The AI revolution’s impact on society, particularly healthcare, is set to surpass the transformative effects of the Industrial and Information Revolutions. Real-world evidence of this can be seen in the application of AI-driven diagnostics and therapeutics, which have significantly enhanced the accuracy and speed of disease detection and rate of targeting and curing disease. The forthcoming role of AI in healthcare is ambitious yet achievable but requires a level of collaboration, innovation, and dedication to a brighter and healthier future.

April 2024

To read the scientific addendum to this whitepaper, click here.

The journey from a research lab to a patient for a single drug spans an average of 12 years. The accompanying cost is by some estimates over $2 billion. Even still, the odds of success remain minuscule, standing at one in 5,000.^1,2 The biotechnology industry, in its current state, appears inefficient; however, the integration of artificial intelligence (“AI”) with exemplary scientific practices presents a promising avenue for significant improvement.

The advent of AI models and applications has sparked a wave of breakthroughs, culminating in the creation of $6 trillion in market capitalization for the “Magnificent 7” stocks alone since the launch of ChatGPT (Figure 1).

Figure 1: The “Magnificent 7” stocks – Microsoft, Amazon, Meta, Apple, Google parent Alphabet, Nvidia and Tesla – drastically outperformed the major U.S. indexes in 2023 as the market recognized them as AI winners. Outperformance YTD has continued with more innovation in the AI space expected in 2024.

Source: Public market data as of April 10, 2024.

Unlike tech stocks, AI’s impact is not yet baked into healthcare valuations. AI has already begun generating new ideas, discoveries and potential opportunities to accelerate the healthcare sector to become more efficient. We can observe this trend by looking at the proliferation of AI mentions in scientific publications and healthcare AI expenditures (Figure 2). The healthcare industry spent approximately $13 billion on AI hardware and software in 2023, with projections to rise to $47 billion by 2028.³ CB Insights reports that between 2019 and 2022, investors injected $31.5 billion in equity funding into AI healthcare ventures.

Figure 2: The number of Pubmed abstracts with the keywords “Machine Learning” or “Artificial Intelligence” has increased dramatically over the past 15 years.

Source: Pubmed.

Broadly on AI: what is it, and why should I care?

AI, in its essence, encompasses the development of computer systems that can perform tasks that typically require human intelligence. These tasks include learning, decision-making, language understanding, and visual perception. The dawn of AI marks a significant leap forward in our ability to process information, automate complex processes and solve intricate problems across almost all sectors and industries. The potential societal impact of AI cannot be overstated. These technologies promise to revolutionize every sector of the economy, from manufacturing to healthcare, education to finance, by enhancing efficiency, unlocking new insights and opening avenues for innovation that were previously unimaginable.

At the core of recent advancements in AI are neural networks inspired by the human brain’s method of processing data inputs and generating predictive outputs. These networks, essential to AI’s evolution, frequently derive their architecture from biological structures, particularly the brain’s complex neural pathways. This biomimetic approach is far from being merely symbolic; it’s fundamentally practical, equipping AI systems to analyze and understand complex information similarly to how biological entities do. Moreover, the deployment of AI in biological contexts provides crucial insights that feedback into AI development, refining algorithms to better emulate natural neural processes. This continuous feedback loop not only drives technological advancements in healthcare but also enhances our comprehension of the biological inspirations behind these AI innovations.

AI’s impact on society and healthcare: shaving billions of dollars and years off the quest to save lives

The AI revolution’s impact on society, particularly healthcare, is set to surpass the transformative effects of the Industrial and Information Revolutions. Real-world evidence of this can be seen in the application of AI-driven models, e.g. Google DeepMind’s AlphaFold, to significantly enhance the speed of drug development.

As with many scientific breakthroughs, Google’s foray into AI within healthcare seems almost serendipitous. Initially, after Google’s DeepMind demonstrated remarkable proficiency in video games, the company decided to leverage its AI capabilities for more academic applications, such as healthcare. This backbone model was used to develop AlphaFold, an AI that predicts a protein’s 3D structure from its amino acid sequences (Figure 3). This is important because there are over 200 million known proteins, each with a unique 3D shape that determines how they work and what they do. Before AlphaFold, scientists could not accurately determine a protein’s unique 3D shape without millions of dollars and years of trial and error. Google taught AlphaFold to predict structures by showing it the sequences and structures of ~100,000 known proteins. Now, AlphaFold has predicted structures for an estimated 98.5% of human proteins with over 92.4% accuracy in positioning atoms, matching experimental methods like X-ray crystallography.² Better yet, AlphaFold’s predictions are freely available to the entire scientific community. This represents a massive leap forward in the quest to generate research tackling disease and finding new medicines. For one, AlphaFold can accurately detect over 50 eye diseases from 3D scans.⁴ This offers a glimpse into a world where early detection and treatment of diseases become the norm rather than the exception.

DeepMind is just one example underscoring AI’s ability to navigate the complexities of biology and healthcare data, translating into life-saving applications and fundamentally changing medical care approaches. Importantly, AI work is just starting; over half of the current 1,500 health AI vendors were established in the last seven years.

Figure 3: Determining protein structure from amino acid sequence.

Source: Baker Lab.

The inherent complexity of biology: the AI revolution will likely play out differently across different areas of healthcare, and data is king

Though Google could (relatively) easily apply its video game AI to biology in just four years, the application of AI in healthcare presents a unique set of challenges, starkly contrasting the relatively straightforward applications seen in other sectors. These challenges primarily stem from the intricate complexity of biological systems and the formidable task of obtaining robust, accurate, and comprehensive data in biology. Unlike the datasets used in finance and marketing, where variables are often well-defined and quantifiable, biological data encompasses many variables that interact in complex, usually unpredictable ways. This intrinsic complexity makes modeling biological systems a daunting task for AI.

One of the foundational hurdles in applying AI to healthcare is the heterogeneity and multidimensionality of biological data. Genetic, environmental and lifestyle factors influence human health and disease, each contributing to the overall complexity of biological systems. For instance, a single genetic mutation can have vastly different effects in different individuals (known as genetic penetrance), depending on the context of thousands of other genes and a myriad of environmental exposures and lifestyle choices. This level of complexity requires AI models to process vast amounts of data and understand and predict the interactions between these variables. This task remains at the frontier of current capabilities.

Moreover, obtaining high-quality, comprehensive biological data poses another significant challenge. Much of the data in healthcare is fragmented across different systems, encoded in unstructured formats or locked within the proprietary databases of research institutions and private companies. Privacy concerns and ethical considerations further complicate data collection, limiting access to the broad datasets necessary for training robust AI models. Even when data is available, it often needs more diversity and representativeness to ensure that AI-driven interventions are effective and equitable across different populations. Even with these limitations, groups like the Chan Zuckerberg Initiative and the Allen Institute are working on collecting robust and well-curved data to help train new models for applications in multiple diseases.

The dynamic nature of biological systems adds another layer of complexity. Unlike static datasets, biological data can change over time, with disease progression, treatment interventions, and lifestyle changes introducing continuous variability. Therefore, AI systems in healthcare must be capable of analyzing static snapshots of data and understanding and predicting how these data points evolve over time, a challenge that requires sophisticated modeling techniques and continuous learning capabilities.

Despite or, better yet, because of these challenges, the potential rewards of applying AI in healthcare are immense. AI has the potential to revolutionize diagnostics, personalized medicine and treatment planning, making healthcare more predictive, personalized and preventive. Overcoming the unique challenges posed by the complexity of biological systems and the difficulty of obtaining robust and accurate data will require innovative approaches to AI model development, data collection and interdisciplinary collaboration between technologists, biologists and clinicians. The journey towards fully realizing AI’s potential in healthcare is complex. Still, the great strides in this direction are promising, signaling a future where AI-driven insights lead to better health outcomes for all.

Great science alone isn’t enough: AI is necessary to continue advancing healthcare

In the dynamic landscape of modern business, adopting AI has transcended from a mere competitive edge to an indispensable necessity. This assertion holds particularly true for the healthcare sector, where AI’s integration promises to refine existing processes and workflows and redefine the paradigms of patient care and medical research. The economic benefits of AI in healthcare are significant; as noted, the healthcare industry is projected to spend $47 billion on AI hardware and software by 2028. This represents a lucrative opportunity for investors and entrepreneurs looking to expand on the AI revolution in healthcare.

AI’s ability to process and analyze data at a scale and speed unattainable by human capabilities alone has led to significant advancements in diagnostics, patient care, drug development and operational efficiencies. For instance, Philips developed the eICU program, which uses AI to aggregate and analyze patient data from ICUs to provide insights into medical decision-making, resulting in a 23% reduction in mortality rates across all four hospitals studied. Furthermore, the average length of stay for a patient in the ICU dropped by 49%, while lengths of stay throughout the health system decreased by 35%.⁵ In drug discovery and development, AI expedites the identification of potential therapeutic targets and predicts the efficacy of drug candidates, dramatically reducing the time and cost associated with bringing new drugs to market.

Beyond these immediate applications, AI facilitates a more personalized approach to healthcare, tailoring treatments and health plans to the individual’s unique genetic makeup and lifestyle factors. This shift towards customized medicine enhances the effectiveness of treatments. It represents a significant leap towards preventive healthcare, potentially saving costs and lives by addressing conditions before they escalate into more severe problems. However, the necessity of AI in healthcare extends beyond improving patient outcomes and operational efficiencies; it is also crucial for the sustainability and growth of businesses within this sector. In an environment where regulatory pressures, rising costs and increasing demands for high-quality patient care converge, AI provides the tools to navigate these challenges effectively. By automating routine tasks, AI allows healthcare professionals to focus on more complex and nuanced aspects of patient care, thereby enhancing patient outcomes. Furthermore, AI’s predictive capabilities enable healthcare organizations to make more informed strategic decisions, from resource allocation to market expansion strategies.

In essence, embracing AI in healthcare is not merely an option for forward-thinking organizations but a strategic imperative for all. Businesses that employ the power of AI will not only lead to innovation and efficiency but will also set the standard for patient care in the future. As AI continues to evolve and its applications within healthcare expand, the gap between adopters and non-adopters will widen, underscoring the urgency for all businesses in this sector to integrate AI into their operations. The go-forward success of healthcare businesses will increasingly depend on their ability to leverage AI technologies to meet the demands of a rapidly changing world, making the case for AI’s necessity not just compelling but categorical.

“Garbage in, Garbage out”: the quality and availability of omics data are critical factors to advancing AI in healthcare

The accuracy and power of AI model predictions, patterns, structures or generative outputs hinge critically on the quality of the input data, bringing the old modeling adage “Garbage in, Garbage out” into sharp relief. While the speed and breadth of applications that AI models have achieved is undeniably impressive, it’s crucial to recognize that the value of data and access to datasets is poised to soon surpass the value of the models. This insight is particularly relevant in the revolution currently underway in healthcare, driven by omics data – genomics, proteomics, metabolomics and more. This revolution is reshaping our understanding of disease, treatment, and wellness, as integrating these omics data types offers a comprehensive view of the biological underpinnings of health, thus paving the way for personalized medicine and targeted therapies.

However, the availability of these omics datasets and methodologies will significantly influence the landscape of healthcare innovation and access. The potential scenarios of a) widespread public availability, b) high restriction or c) a balanced approach to data access each illustrate varying impacts on healthcare progress, underscoring the intertwined future of AI models and omics data in defining the next frontier in healthcare.

• Scenario A: omics data is democratized, likely leading to an unprecedented acceleration in healthcare innovation – With scientists, researchers and clinicians worldwide having open access to vast datasets, the collaborative potential for discovering new disease mechanisms, therapeutic targets and treatment modalities would expand exponentially. This openness could also foster a more inclusive research environment, where even institutions with limited resources could contribute to and benefit from global scientific advancements. This is the type of open collaboration fostered by AlphaFold. It was built on publicly available datasets, and the AI model itself was made publicly accessible for anyone to use. Moreover, initiatives have been undertaken to standardize and curate massive genomic datasets, such as those by The Cancer Genome Atlas (“TCGA”), and the UK Biobank Proteomics and other datasets. All of these efforts have taken decades – and the dedication of tens of thousands of researchers – and still require massive curation and maintenance to generate effective AI model outputs.
• Scenario B: omics data access is tightly controlled by select entities, potentially leading the deployment of AI to become a double-edged sword – On one hand, AI could streamline research and development within these privileged circles, pushing the boundaries of proprietary medical breakthroughs. On the other hand, this restriction could limit AI’s learning potential by confining it to siloed datasets, potentially slowing the pace of innovation and widening the gap in healthcare access. Given the value of these datasets, we have already seen private entities begin to hoard data for commercial purposes or continue to restrict access to proprietary datasets to maintain a competitive edge. Companies like Flatiron Health and Tempus  have been gathering massive datasets for commercialization purposes, while almost all therapeutic companies restrict the publication of their internal datasets to maintain trade secrets and competitive advantage.
• Scenario C: a middle-ground, balanced approach to omics data availability has the highest likelihood of unfolding in the near future – There will likely be significant academic and government-funded generation and curation of large omic datasets, with private companies restricting access to proprietary datasets. Over time, these large public datasets will mature and expand to include new omics datasets, and standardization will likely play a crucial role. Large initiatives like the Chan Zuckerberg Initiative are already building databases with the scale and infrastructure necessary to generate high-quality datasets. Private companies will likely release data after it has been depleted of value or when there is greater value in sharing the data than in hoarding it. We have already seen small biotechs like Leash Bio adopt an open-source approach and share data with the strategy that more integration and data will only enhance their efforts to develop novel therapeutics.

Venture capital’s role in driving Healthcare x AI forward

From the vantage point of investment in the rapidly evolving landscape of healthcare, the integration of AI stands out not just as an opportunity but as a critical imperative. Investors increasingly recognize that the future healthcare leaders will be those organizations that adeptly harness AI to drive innovations, improve patient outcomes, and streamline operations. In contrast, companies that need to be faster to adopt AI technologies risk falling behind and potentially facing obsolescence. The stark reality is that the healthcare sector is at a pivotal inflection point where the adoption of AI can significantly differentiate between market leaders and laggards. These dynamic underscores the critical importance of investment strategies that prioritize and support the integration of AI in healthcare.

The argument for prioritizing investments in AI-enhanced healthcare companies is multifaceted. Firstly, AI’s ability to process and analyze vast datasets at speeds and accuracies unattainable by human capability transforms diagnostics, treatment personalization, and patient monitoring. This technological prowess is pivotal in decoding complex health data, leading to breakthroughs in understanding disease mechanisms and developing targeted therapies. For investors, backing companies that leverage AI in these areas offers the promise of substantial financial returns and the opportunity to contribute to significant societal impacts in terms of improved health outcomes and longevity.

Furthermore, the synergy between great science and AI utilization forms a cornerstone for the next generation of healthcare innovation. AI does not operate in a vacuum; its most significant potential is unlocked when combined with cutting-edge scientific discoveries and insights. Companies that cultivate a culture of innovation, where AI tools are used to accelerate and expand the impact of scientific research, are poised for success. This fusion of AI with scientific inquiry creates new frontiers in personalized medicine, making treatments more effective and reducing adverse outcomes. From an investment perspective, this represents fertile ground for growth, as companies at the forefront of this convergence are likely to be in charge of introducing novel healthcare solutions and capture significant market share.

Investors, including us at Luma Group, are encouraged to scrutinize the current AI capabilities of healthcare companies and their potential to innovate and adapt. The criteria for investment decisions should include an evaluation of a company’s commitment to AI integration, the robustness of its data infrastructure, the caliber of its scientific research, and its capacity for leveraging AI to translate scientific discoveries into clinical applications. Companies that demonstrate a clear vision for integrating AI with scientific exploration are the ones that will navigate the complexities of modern healthcare, delivering solutions that address unmet medical needs and driving the future of medicine.

One of our key focus areas here at Luma Group is investing in AI-driven enterprises pioneering drug discovery and diagnostics advancements. Among our investments is Rome Therapeutics, which delves into large human genetics datasets to explore the “dark genome”—a segment historically dismissed as evolutionary junk, with no coding for proteins or known biological function. Rome Therapeutics has discovered novel therapeutic targets for autoimmune diseases hidden within these previously overlooked genomic territories through its innovative AI models. Their AI approach examined the vast dark genome region to identify regions highly expressed in patients with specific autoimmune diseases. Within these regions, they discovered nonconical protein expression patterns and found ancient retroviral proteins that are driving the disease progression. This breakthrough could not have been achieved without the sophisticated analysis enabled by their AI technology.

Our recent investment Curve Biosciences is a trailblazer in diagnostics focusing on developing cutting-edge early-stage liver diagnostics that will leverage AI. Curve’s data engine analyzes methylation patterns in cell-free circulating DNA in the blood. While it is understood that total methylation levels escalate over time, the exact tissue methylation patterns distinguishable in cell-free DNA that strongly relate with disease progression have remained elusive. Through the application of AI to analyze vast patient tissue DNA methylation datasets across various stages of chronic disease progression, Curve Biosciences has pinpointed predictive methylation signatures with subtle, previously undetectable patterns of methylation that strongly correlate with the stage of and progression of the disease (Figure 4). This innovative approach exemplifies the power of AI in transforming diagnostics and how we understand disease mechanisms with an immediate application in diagnostics.

Figure 4: Curve leverage AI/ML to process large PCR, microarray, and sequencing data sets to inform their screening technology.

Source: Company Materials.

The investment perspective on AI in healthcare is clear: leveraging AI is not merely an option but a strategic imperative for success. As the sector continues to evolve, the combination of innovative science and AI utilization will distinguish the leaders from the followers. For investors, the message is equally clear: support and invest in companies that understand and act on this imperative, for they are the ones that will shape the future of healthcare, delivering both substantial financial returns and transformative improvements in human health outcomes.

Call to action

In the vast and ever-evolving landscape of healthcare, where the promise of groundbreaking innovation stands on the horizon, a profound opportunity exists to reshape healthcare and make meaningful improvements in human health. The forthcoming role of AI in healthcare is ambitious yet achievable but requires a level of collaboration, innovation, and dedication to a brighter and healthier future.

Looking ahead, AI’s influence on healthcare will only broaden as the technology itself advances. At some point in the not-so-distant future, we anticipate that the advent of Advanced General Intelligence (“AGI”) could lead to a whole new era in healthcare by surpassing human intelligence to predict diseases, tailor treatments to individual genetics and automate complex medical procedures. AGI promises to enhance diagnostic accuracy, optimize healthcare management and improve resource allocation, fundamentally increasing the quality and accessibility of care. AI’s path forward in this space will be complex and filled with challenges, yet it is paved with the potential for extraordinary advancements. This will require a collective effort by researchers, clinicians, technologists, entrepreneurs and investors, sharing knowledge, resources and visions to catalyze the development of solutions that can address the most pressing health challenges of our time. Moreover, this collective endeavor demands a commitment to relentless ideation and innovation. We are at the cusp of a new era in healthcare that leverages AI, genomics and emerging technologies to make new insights into human health and leverage them for new therapeutics, clinical trials, diagnostics and other healthcare applications. We believe AI will dramatically improve patient outcomes and redefine healthcare in the

---

1. https://www.pwc.com/gx/en/industries/healthcare/publications/ai-robotics-new-health/transforming-healthcare.html ↩︎
2. https://www.biopharmadive.com/news/new-drug-cost-research-development-market-jama-study/573381/ ↩︎
3. https://www.economist.com/technology-quarterly/2024/03/27/ais-will-make-health-care-safer-and-better ↩︎
4. https://deepmind.google/discover/blog/a-major-milestone-for-the-treatment-of-eye-disease/ ↩︎
5. https://www.fiercebiotech.com/medtech/philips-illustrates-gains-over-15-years-a-telehealth-powered-icu ↩︎

Advances in drug development and discovery have revolutionized the delivery of life-saving treatments to patients. Despite the immense impact of these innovations, the journey from concept to bedside remains largely misunderstood. Through this primer series, we explore the complex processes behind drug discovery, highlighting the challenges the industry still faces and the critical need for investment of time and resources to deliver the next generation of medicines.

September 2024

Introduction

Luma Group’s mission is to invest in the most innovative companies poised to make meaningful impacts on human health outcomes. We focus on therapeutics, diagnostics, and medical devices because they provide a unique opportunity for venture capital to thrive, while fueling innovation that will benefit humanity. Despite the enormous impact that medical improvements have had on nearly all lives, the process to develop these drugs (”drug discovery”) remains poorly understood by most due to its long, arduous, expensive and highly complex nature.

We seek to shed light on this topic, and several others in our primer series. Through these primers, we hope to provide additional clarity into several integral processes underlying the biotech industry to help a broader understanding of the challenges the industry faces and the necessity for continued investing here. We hope readers walk away with a deeper appreciation and a clearer understanding of the immense and collaborative effort involved in the biotechnology community’s journey to advance drug discovery and development. It truly takes a village to create a drug.

In part one of this primer series, we will explore the various steps of the drug discovery process, delving into the key elements of each step and the challenges and opportunities for future growth (Figure 1). Despite the inherent complexities and obstacles, we are in the golden age of innovation within our sector. Our goal is to help demystify the journey of how drugs progress from the laboratory bench to the patient’s bedside.

In this document, we break down the drug discovery process into distinct steps and summarize each with a table that reviews key aspects. These aspects include the step’s “significance,” which refers to its importance in the overall process of creating a drug; its “purpose,” highlighting the key data generation objectives; its “probability of in-stage success (PoS),” indicating the likelihood that a drug candidate will successfully advance to the next stage; its “timelines/costs,” which detail the expected duration and associated expenses; and its “areas for optimization,” where we identify opportunities for companies to enhance their risk-adjusted outcomes in drug discovery.

To supplement this series and other Luma content, we have compiled a target glossary of key technical terminology to support readers. This glossary can be accessed here.

For every 100 candidates entering the drug discovery process, it is estimated that only between 1 to 3 will ultimately become FDA-approved drugs that reach commercialization. However, it is important to note that in the earliest stages of development, such as compound screening, depending on the methodology used, a group may need to screen tens of thousands of compounds just to find a single hit. This reality underscores the challenge of quantifying asset attrition, and the graphic below serves to illustrate the monumental undertaking required to get a new drug to patients.

Figure 1:  Overview of critical endpoints, cost, and risk profile of each discovery and development stages.

Asset counts beings after developmental candidate nomination as after this point a drug format has been finalized for development and failure becomes a binary process. Source: Deal Room, Speedinvest.

Importance of Drug Discovery

Modern medicine has revolutionized billions of lives, from using antibiotics to combat once-deadly bacterial infections to developing cures for life-threatening genetic diseases. Although only a few drugs ultimately reach the hands of physicians, countless hours and numerous resources are poured into these efforts of discovering and developing the next generation of life-saving drugs. The entire journey – from taking a new biological discovery in a university lab to eventually saving a patient’s life – is what we think of as “drug discovery”.

Figure 2:  Key steps in the drug development process.

Source: Luma proprietary.

Early pioneers in drug discovery laid the foundation for our modern drug discovery process. Visionaries like Louis Pasteur, who pioneered the germ theory of disease, Alexander Fleming, who discovered penicillin, Jonas Salk, who developed the polio vaccine, and Dorothy Hodgkin, who unlocked the structures of important biochemical substances, dedicated their lives to transforming their scientific breakthroughs into what we now recognize as life-saving drugs.

Their work has inspired countless researchers to convert their own insights into new therapies that continue to save lives. In today’s landscape, regulators play a crucial role in drug discovery, and understanding their influence is essential to grasping the complexities of the process.

Over the past 120 years, the FDA in the US and the EMA in the EU have standardized the drug approval process, ensuring that only the safest drugs and most effective drugs reach the market. The rigorous standards set by these regulatory bodies have established best practices across the industry, even influencing other jurisdictions, thereby accelerating the timeline from discovery to approval and significantly increasing the efficiency of approvals. However, while this robust regulatory framework has positive outcomes, it is lengthy and costly, typically requiring $350 million and 10-15 years to bring a promising discovery to the point of commercial (i.e., patient care).¹  Navigating this complex process demands both experience and substantial capital, critical resources provided by venture capital investors to the innovative scientists driving discovery.

The global drug market has seen remarkable and sustained growth over the last 50 years, expanding from around $12 billion in the 1970s to over $1.56 trillion by 2020.² This tremendous growth, driven by the development of new therapies and the expansion of global markets, is expected to continue at approximate 6% compound annual growth rates, potentially reaching nearly $3 trillion within the next decade.³

This growth is driven by advancements in biotechnology, increased investment in research and development and the introduction of innovative therapies for chronic and complex diseases. For instance – the development of biologics, such as monoclonal antibodies and gene therapies, has revolutionized treatment options for conditions like cancer, rheumatoid arthritis, and genetic disorders.

The success of drugs such as Humira, one of the world’s best-selling drug (over ~$20B in sales/year), illustrates the potential for growth and impact in the pharmaceutical industry. Additionally, the rapid development and deployment of COVID-19 vaccines demonstrated the industry’s and regulator’s capacity for swift innovation and large-scale production in response to global health crises.

The Long Journey of Drug Discovery and Development

Step 1. Target Identification and Validation in Drug Discovery

Target identification and validation is the first step in the extensive drug discovery and development process. This crucial phase involves selecting a specific node within a disease pathway to target with a new drug. Typically, this process originates from foundational research conducted in academic labs, where scientists strive to understand the underlying biology of specific diseases. For example, researchers may find a protein or molecule crucial to disease progression (the “target”).

This foundational knowledge is the springboard for drug development – the earliest stage. Successfully transitioning this early-stage academic research into a plausible biotech venture often hinges on validating the target – expanding on the academic research that’s already been completed. See “Validating Drug Targets” below for a further explanation.

Identifying Drug Targets

Identification can take a top-down (“clinical sample”) or bottom-up (“research sample”) approach for identification. The top-down approach typically looks for causal relationships of a target from a patient sample and requires basic research to validate the causal nature. For instance, in breast cancer research, scientists took a clinical top-down approach to identify the HER2 protein, which is overexpressed in a subset of breast cancer tumors. This discovery was pivotal as it led to the development of targeted therapies like trastuzumab (Herceptin), which specifically inhibits HER2, thereby slowing tumor growth and improving patient outcomes.

The bottom-up approach aims to perturb specific target biology and look for outcomes reflecting patient disease pathology (Figure 3). Paclitaxel, a pioneering chemotherapeutic agent, was discovered through natural product research. Scientists at the US National Cancer Institute identified a compound in the Pacific yew tree (Taxus brevifolia) that inhibited cancer cell growth in cellular models. Later synthesized in laboratories to enhance production efficiency, this discovery has profoundly advanced the treatment of several cancers, including ovarian and breast cancer.⁴

The clinical sample approach has progressively become the first choice in drug development. By starting with patient samples from the target disease, it can help researchers identify mutations or aberrations only present in the diseased patients and provide a narrowed starting universe for testing. In the bottoms-up approach, discovery relies more heavily on trial and error as there isn’t as robust of an underlying hypothesis around potential disease drivers. However, once the target has been identified and validated, both paths converge on the same output to be taken forward into further stages of development.

Figure 3: Target identification and characterization workflow.

Source: Nature.

Validating Drug Targets

Once a potential target is identified, its role in the disease must be validated to ensure it is a viable candidate for drug development. This is where advanced techniques like “knockout mice” and CRISPR technology come into play.

Knockout mice are mice that are genetically engineered to lack a specific gene – generally the gene associated with the “target,” allowing researchers to observe the effects of its absence on disease progression. This method both offers more data to evaluate the target and helps to establish whether targeting the relevant gene can alter the disease’s course.

CRISPR technology has revolutionized target validation by enabling precise gene editing. This pioneering work has made it significantly easier and cheaper to manipulate multiple genes in cellular and animal models. Researchers can selectively disrupt or modify genes in cell lines or animal models to study their function and contribution to disease. This precise control can provide compelling evidence of a target’s relevance and potential as a therapeutic focus.

Real-World Example: Bridging Academia and Industry

To progress in the drug discovery process, an academic team generally needs to be brought into a commercial environment for capital formation. Spinning out academic research into the pharmaceutical industry requires strategic partnerships, and universities and research institutions often collaborate with biotech companies beginning at this stage. The large academic centers maintain specialized staff who are experienced in creating these collaborations in a way that rewards the university (or other institution) to the extent the commercialization effort is successful. A large part of our opportunities at Luma Group come from this nexus.

Key Takeaways for Step I

Step 2. Hit Identification and Screening in Drug Discovery

The next pivotal phase in drug discovery after target identification and validation is hit identification and screening. This stage involves finding compounds that can interact with the identified target, which may lead to a potential therapeutic effect. This step typically involves screening thousands to billions of compounds to identify the roughly <0.5% that interact with the target. The goal is to identify those that show some level of activity – these are referred to as “hits” and require further refinement and optimization. This is an iterative cycle, where researchers use the early results to increase the chances of finding more hits. Some of these hits will be taken forward as candidates and some may serve as backups or future “next generation” assets against the same target. Different strategies can be employed in hit identification and screening, which can provide unique advantages or differentiation in the drug discovery process. We will only cover the most common strategies in this primer.

High-Throughput Screening (HTS)

High-Throughput Screening (HTS) is a crucial and powerful tool in drug discovery that leverages automation to rapidly screen large libraries of compounds – ranging from thousands to millions – against a specific biological target. By utilizing advanced robotic systems and automated processes, HTS allows for the simultaneous execution of numerous biochemical, genetic or pharmacological assays, significantly accelerating the identification of active compounds, antibodies or genes that influence particular biomolecular pathways. One of the key advantages of HTS is its effectiveness in identifying inhibitors of kinases – a class of enzymatic proteins that play pivotal roles in regulating various cellular processes. Given their importance in disease pathways, kinases are prime targets for HTS, which can efficiently sift through extensive compound libraries to pinpoint molecules that inhibit these critical proteins. These libraries, which may be commercially available, proprietary, or a combination of both, typically contain thousands to billions of compounds.

Despite its broad utility, HTS has limitations.  It is less effective when applied to smaller compound libraries, or when the target has limited biological reagents, such as cell-free recombinant proteins, or when structural information about the target is minimal. Despite these challenges, HTS remains a cornerstone of the drug discovery process by identifying promising hits for further development.

Alternative Approaches to Hit Identification

While HTS remains a fundamental method in modern drug discovery, alternative approaches are often necessary, especially when HTS is either impractical or inefficient. For target proteins with poorly understood biology or those from challenging non-kinase classes (such as transcription factors and cell membrane proteins), these alternative strategies may be more successful in generating early-stage drug-like compounds.

1. Fragment-Based Drug Discovery (FBDD): Fragment-Based Drug Design involves identifying small chemical fragments that bind to a target protein using a hybrid HTS approach that leverages protein structural information. These fragments are typically simpler than traditional drug molecules, making analyzing their interactions with the target easier. Once a fragment is identified, medicinal chemists build on it, adding functional groups to enhance its binding affinity and specificity. While FBDD can address some of the limitations of traditional HTS, developing fragments into drug-like scaffolds requires significantly more effort. This process demands a substantial amount of structural information to guide the design and optimization. Despite these challenges, FBDD enables the creation of highly optimized drug candidates from small, initially weak-binding fragments.
2. Virtual Screening: Virtual screening leverages advanced computer models to predict which compounds are more likely to bind to a target, thereby eliminating the need for labor-intensive wet-lab based HTS methods. This approach simulates interactions between the target and potential drug molecules in silico, enabling researchers to prioritize compounds before any experimental testing. By exploring much larger compound libraries and focusing on the most promising candidates, virtual screening significantly reduces the time and cost associated with hit identification, streamlining the process for subsequent wet lab testing.

Although this approach can save much time and effort compared to manual HTS, it generally yields less reliable results. This is mainly due to limitations inherent in in silico approaches, such as a lack of atomic-level detail, unpredictable outcomes and challenges associated with modeling complex biological systems. While virtual screening is a powerful tool, it is regarded more as a supplementary method than a primary one. It will require further advancements before transitioning to a reliable and widely used primary screening method.

Real-World Example: Discovery of Gleevec (Imatinib)

A notable success story in hit identification through HTS is the discovery of Gleevec (imatinib), a groundbreaking treatment for chronic myeloid leukemia (CML). Researchers identified the BCR-ABL fusion protein as a key driver of CML. Using HTS, they screened hundreds of thousands of compounds to find inhibitors of this protein.⁵ Imatinib emerged as a potent and specific inhibitor of BCR-ABL, leading to its development and subsequent approval by the FDA. Gleevec has transformed CML from a fatal disease into a manageable condition for many patients, showcasing the power of HTS in drug discovery.

Key Takeaways for Stage II

Step 3. Lead Discovery and Optimization in Drug Development

The next critical phase is lead discovery and optimization after identifying potential hits in the drug discovery process. This stage focuses on picking one or more candidate compounds and then refining and improving those compounds to meet specific criteria, ensuring they possess the necessary properties to become effective drugs. Lead compounds are characterized by their potency, selectivity, and overall drug-like properties (see “Criteria for Lead Compounds,” below). This critical step can be one of the longer steps in the process, as development teams make minor modifications to the compound or compounds and repeatedly test those changes to develop stronger candidates for further development (figure 4). It would not be unusual for a team to test hundreds – or even thousands – of potential modifications before settling on a lead candidate.

Criteria for Lead Compounds

Optimization Strategies

Once lead compounds are identified, they undergo extensive optimization to enhance their properties and maximize their therapeutic potential. This crucial step in drug development involves refining the compounds’ efficacy, safety, and pharmacokinetic properties. While minor optimizations may occur later steps, this initial optimization phase is essential for defining the core characteristics and bulk properties of the drug.

Structure-Activity Relationship (SAR) Studies: SAR studies involve systematically modifying the chemical structure of lead compounds to understand the relationship between their structure and biological activity. By making small changes to the molecule, researchers can determine which modifications improve potency, selectivity, or other desirable characteristics. This iterative process helps fine-tune the compound to achieve optimal performance.

Medicinal Chemistry: Medicinal chemists play a crucial role in optimizing lead compounds. They use their expertise to design and synthesize new analogs, aiming to improve bioavailability, reduce toxicity, and enhance overall drug-like properties. This often involves addressing issues such as poor solubility, metabolic stability, or undesirable pharmacokinetics. Through medicinal chemistry, compounds can be modified to ensure they are suitable for further development.

Figure 4: Workflow for the iterative lead optimization process.

Source: Nature.

Real-World Example: Optimization of HIV Protease Inhibitors

A prime example of successful lead optimization is the development of HIV protease inhibitors. Initially, researchers identified compounds that could inhibit the HIV protease enzyme, which is essential for the maturation and replication of the virus. However, early inhibitors had limitations in terms of potency, bioavailability, and side effects. Through extensive SAR studies and medicinal chemistry efforts, scientists were able to optimize these inhibitors. They made structural modifications in the compound’s p2 and P2” region that binds in the enzymatic domain of the HIV proteases, enhancing potency against drug-resistant variants​.⁶ Additionally, they addressed issues related to bioavailability and metabolic stability, resulting in compounds that were more effective and had fewer side effects. For instance, the optimization of saquinavir, the first HIV protease inhibitor approved by the FDA, involved modifying its chemical structure to improve its pharmacokinetic profile. Researchers at Roche achieved these enhancements using computer-aided SAR and structure-based drug design.⁷ Subsequent protease inhibitors, such as ritonavir and lopinavir, were further refined to enhance efficacy and reduce adverse effects, significantly improving the treatment outcomes for patients with HIV/AIDS​.

Key Takeaways for Step III

Step 4. IND Enabling Studies

After the lead optimization phase, promising drug candidates enter preclinical development. This critical stage involves extensive testing to evaluate the drug candidate’s safety, efficacy, and pharmacokinetic properties before it progresses to clinical trials. IND-enabling work includes both in vitro and in vivo studies, providing a comprehensive understanding of the drug’s potential effects on biological systems. Though many of the approaches are similar to those used in lead optimization, they are typically much larger and longer studies done in cells and animals to understand the behavior of the drug holistically. These studies form the foundation for a data package submission to the FDA. FDA approval is required before a candidate can begin human testing in the US. Other countries have similar standards (Figure 5).

In Vitro Studies

1. Testing on Cell Cultures: In vitro studies involve testing drug candidates on cell cultures to assess their efficacy and toxicity. Researchers use various types of cells, including human cells, to observe how the drug interacts with its target and to identify any cytotoxic effects. This step is crucial for determining the drug’s initial efficacy and for screening out compounds with high toxicity levels early in the process.
2. ADME Studies: Absorption, Distribution, Metabolism, and Excretion (“ADME”) studies are essential in understanding how the drug behaves in the body. These studies help predict the drug’s bioavailability, distribution throughout the body, metabolic pathways, and excretion. ADME profiling ensures that the drug reaches its target in sufficient concentrations and has an appropriate half-life, minimizing the risk of adverse effects.

In Vivo Studies

In vivo, studies involve testing drug candidates in animal models to evaluate their safety and efficacy in living organism live mammals spanning from mice to non-human primates (new world monkeys). These studies provide critical insights that cannot be obtained from in vitro experiments alone. For example, dog and rat models are commonly used to understand the safety profile and how the drug is metabolized in the body. These models help researchers understand how the drug affects the entire organism and identify potential side effects.

Real-World Example: IND Enabling Testing of Monoclonal Antibodies

A notable real-world example of IND-enabling development is the testing of infliximab, a monoclonal antibody for autoimmune diseases. Engineered to target specific antigens involved in disease processes, infliximab underwent rigorous preclinical testing to ensure its safety and efficacy before reaching clinical trials. Cell-based in vitro studies confirmed its ability to neutralize tumor necrosis factor-alpha (TNF-α), a cytokine involved in inflammatory responses. Researchers developed a predictive mouse model of arthritis and tested whether these arthritic mice would show slowed or halted disease progression when treated with infliximab​.⁸ These in vivo studies demonstrated infliximab’s efficacy in reducing inflammation and preventing disease progression without significant toxicity​.⁹ These animal studies showed that infliximab could effectively reduce symptoms and inflammation markers, providing clear evidence of its therapeutic potential. The robust preclinical results supported advancing infliximab into clinical trials. This eventually led to its approval and widespread use in treating various autoimmune diseases.

Figure 5: Overview of the components for an IND Data Package.

Source: Adgy Life Sciences.

Key Takeaways for Step IV

Clinical Development will be covered in Part 2 of the primer series

Current Challenges in Drug Discovery

High Attrition Rates: One of the most significant challenges in drug discovery is the high attrition rate. The journey from initial discovery to an approved drug is fraught with obstacles, and many promising candidates fail during development. This failure can occur at any stage, from preclinical testing to late-stage clinical trials, often due to unforeseen toxicity, lack of efficacy, or poor pharmacokinetics. Even though many of the individual stages have high percentages, the communicative attrition rates result in>30% of drug candidates that enter clinical, and only 5-10% of these drug trials ultimately receive FDA approval. Still, depending on the stage at which clinical challenges emerge, the relevant company could have spent hundreds of millions of dollars.¹⁰ It may require further investment to remediate the challenges, or the endeavor could fail altogether. This means that any successful product’s commercial value must be high enough to offset the cost of any failures. 

High Cost of Research: Over the past 75 years, drug development costs have continuously escalated. This phenomenon, humorously dubbed Eroom’s Law, reflects an inverse trend to Moore’s Law for microchip development.¹¹ Eroom’s Law observes that the cost of drug development roughly doubles every year, in contrast to Moore’s Law, which states [that the number of computations a semiconductor can complete, per unit, etc etc.]. As a result, the barriers companies face in bringing a drug to market has continued to escalate. All else equal, we view this as one of the factors that continues to support high average returns on capital to the biotech venture capital industry; demand for capital continues to increase at a time in which this specialized form of capital remains scarce.

Figure 6: Eroom’s Law depicted by illustrative historical regression of significant increase of drug discovery and development.

Source: Benchling.

Valley of Death:  Significant requirements for capital and time asymmetries exist between value generation and risk reduction in drug development. Combined with the high costs and binary nature of outcomes, this necessitates that companies raise or spend substantial capital to mitigate risks during the development process. These gaps contribute to the “valley of death,” a critical phase where companies may lack the necessary funds to navigate lengthy and expensive stages of development, potentially forcing them to halt their programs (Figure 7). For instance, the biotech company Epizyme faced considerable financial challenges in advancing its cancer drug, Tazemetostat, through clinical trials but overcame these issues to be approved and is now commercially available. Despite promising early results, the high costs and extended timelines required significant capital, leading to delays and funding gaps that exemplify the difficulties of the “valley of death” in drug development.

Figure 7: Illustration of the drug discovery valley of death.

Source: Research gate.

Future Directions in Drug Discovery

Artificial Intelligence (AI): AI is revolutionizing drug discovery by enhancing the ability to predict, analyze and understand complex biological data. The integration of AI in drug discovery has the potential to significantly reduce the time and cost involved in bringing new drugs to market. Here are several ways AI is transforming this field:

1. Predicting Protein Structures: One of the most notable applications of AI in drug discovery is predicting protein structures. Accurate prediction of protein folding is crucial for understanding how drugs interact with their targets. DeepMind’s AlphaFold has made significant breakthroughs in this area. AlphaFold’s ability to predict the three-dimensional structures of proteins with high accuracy has provided researchers with vital insights into the fundamental mechanisms of diseases. This advancement accelerates the identification of new drug targets and designing molecules that can interact precisely with these targets.
2. Drug Design and Optimization: AI algorithms are being used to design and optimize drug candidates. Machine learning models can analyze large datasets of chemical compounds and their biological activities to predict which compounds will have the desired effects on specific targets. This process can save millions of dollars and accelerate the process in the lead optimization step by year. For example, Insilico Medicine utilized AI to identify new drug candidates for fibrosis, significantly speeding up the discovery process. AI-driven platforms can generate novel compounds, optimize existing ones, and predict their pharmacokinetic and toxicological properties, streamlining the drug design phase.
3. Virtual Screening: AI enhances initial screening by virtually predicting which compounds are most likely to bind to a target based in the hit to lead stage. Traditional virtual screening methods can be time-consuming and computationally intensive, but AI can process vast libraries of compounds more efficiently. Companies like Atomwise use AI to conduct initial virtual screenings, identifying potential drug candidates in a fraction of the time required by conventional methods. This approach increases the likelihood of finding effective drugs and reduces the reliance on high-throughput physical screening.
4. Biomarker Discovery: Identifying biomarkers is essential for developing targeted therapies and companion diagnostics. These biomarkers are samples from the patient that can help understand the disease progression and how well the patient is reacting to therapy. AI can analyze complex biological data, such as genomics, proteomics, and metabolomics, to identify potential biomarkers that indicate how a patient will respond to a particular treatment. IBM Watson for Drug Discovery, for example, uses AI to uncover hidden connections in biomedical data, aiding in the discovery of new biomarkers and therapeutic targets.
5. Predicting Clinical Trial Outcomes: AI models can predict the outcomes of clinical trials by analyzing historical trial data and identifying factors that influence success rates. This predictive capability helps in designing more effective trials, selecting the right patient populations, and optimizing protocols. AI-driven predictive analytics can reduce the risk of trial failure, saving time and resources.

We recently published a whitepaper on the impact of AI on healthcare, which you can read here.

Novel Therapeutic Modalities: The development of novel therapeutic modalities, including gene therapy, CRISPR and biologics, is opening new frontiers in medicine. Gene therapy aims to treat or prevent diseases by introducing, removing, or altering genetic material within a patient’s cells. An example is the FDA-approved gene therapy Luxturna, which treats a rare inherited vision loss. CRISPR technology, which allows precise genome editing, holds promise for treating genetic disorders by correcting mutations at their source. Biologics, such as monoclonal antibodies and cell-based therapies, offer targeted treatment options for a wide range of diseases, including autoimmune disorders and cancers.

Real-World Examples

AI in Drug Discovery: Exscientia was an AI-driven drug discovery company. In 2020, Exscientia collaborated with Sumitomo Dainippon Pharma to develop DSP-1181, a drug candidate for treating obsessive-compulsive disorder (OCD). Using its AI platform, Exscientia was able to design the drug in just 12 months, a process that typically takes 3 to 5 years using traditional methods. The AI system efficiently analyzed and processed vast amounts of data to identify and optimize the drug candidate, demonstrating a significant acceleration in the drug discovery process. Specifically, they used AI to rapidly predict new chemical structures during lead optimization, significantly faster than the traditional method DSP-1181 advanced to clinical trials, highlighting the potential of AI to speed up the development of new medications.¹² Exscientia entered an agreement to be acquired by Recursion Pharmaceuticals in August 2024.

CRISPR and Gene Therapy: The application of CRISPR technology in clinical settings is exemplified by ongoing trials for sickle cell disease and beta-thalassemia. Companies like CRISPR Therapeutics and Vertex Pharmaceuticals are leading the way with CRISPR-based treatments that aim to edit the faulty genes responsible for these conditions. Early results have shown promise, with some patients achieving significant improvements, highlighting the potential of gene editing as a therapeutic modality.

Precision Autoimmune Diseases: The development of biological therapies, such as adalimumab (Humira) and infliximab (Remicade), has pioneered the treatment of autoimmune diseases like rheumatoid arthritis and Crohn’s disease. These drugs specifically target inflammatory pathways, offering relief to patients who do not respond to traditional therapies. A significant challenge in treating autoimmune conditions is the ability to precisely target disease cells or tissues while avoiding toxicity in healthy ones. Recent biological insights have enabled the adaptation of clinically derisked technology from precision oncology for autoimmune applications. Luma recently published a white paper highlighting advancements and opportunities in this innovative field. The success of these biologics underscores the importance of targeted treatments in managing complex chronic conditions.

Concluding Thoughts

Drug discovery is not for the faint of heart. It requires unwavering conviction, passion, perseverance, a vast team of dedicated individuals, and substantial financial resources to bring a therapy to market. Despite these challenges, countless people in our sector commit decades, even entire lifetimes, to the hope of developing a drug that can offer patients a better life than their disease allows today. Although the journey can be daunting, it is profoundly rewarding to know that all this effort, time, and investment can significantly impact patients’ lives and improve global health within our lifetimes.

---

1. https://www.forbes.com/sites/matthewherper/2013/08/11/how-the-staggering-cost-of-inventing-new-drugs-is-shaping-the-future-of-medicine/#7753cc096bfc ↩︎
2. https://www.biospace.com/article/releases/pharmaceutical-market-size-to-hit-around-usd-2-832-66-bn-by-2033/ ↩︎
3. https://www.biospace.com/article/releases/pharmaceutical-market-size-to-hit-around-usd-2-832-66-bn-by-2033/ ↩︎
4. https://web.archive.org/web/20150905144824/https://dtp.nci.nih.gov/timeline/flash/success_stories/S2_Taxol.htm ↩︎
5. https://www.nature.com/scitable/topicpage/gleevec-the-breakthrough-in-cancer-treatment-565/ ↩︎
6. https://www.mdpi.com/1422-0067/23/22/14178 ↩︎
7. https://link.springer.com/chapter/10.1007/978-981-16-5180-9_16 ↩︎
8. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC453150/ ↩︎
9. https://www.sciencedirect.com/science/article/abs/pii/016158909390106L ↩︎
10. https://link.springer.com/article/10.1007/s43441-023-00509-1 ↩︎
11. https://www.nature.com/articles/nrd3681 ↩︎
12. https://www.sumitomo-pharma.com/news/20200130.html ↩︎