To kick off the New Year, we spoke with leaders and experts in the Bio-IT community about their predictions for 2024. AI and machine learning advancement seems to be at the forefront of everyone’s mind as it continues to develop new and innovative uses. Despite some concerns about AI’s impact on the job market, Ariel Katz of H1, predicts, “Pharmaceutical companies are on the brink of a tech revolution, and will embrace AI as a means to enhance their operations and research capabilities.”
AI’s advancement will bring changes in data management, organization, and sharing. “Pharma will keep closing the digital transformation gap,” predicts Alister Campbell of Dotmatics. “Life science companies will start to see the fruits of digitalization as they increase the use of drug repurposing, making it easier to navigate the swaths of data that have previously been locked away in silos, unlocking the potential of drugs that have been shown to be safe but missed initial clinical endpoints.”
Finally, gene editing and CRISPR technologies are anticipated to bring cell and gene therapies “into a new era of precision and complexity,” according to Michelle Fraser of Revvity. “Developers in the cell and gene therapy arena are fervently working to create improved therapies characterized by greater efficacy, improved safety profiles, and streamlined manufacturing processes. At the heart of this pursuit lies innovations in payload and delivery mechanisms. Within the gene editing landscape, developers are narrowing in on novel nucleases with specificity to, and optimal activity, at the targeted site.”
Here are the full trends and predictions, including more information on AI and machine learning advancements, spatial omics, data management, and gene editing. –the Editors
Andreas Busch, Chief Innovation Officer, Absci
Better data drives quicker drug development—especially AI-powered data. In 2024 we’ll see a surge of partnerships to put more data to work. Drug researchers will expand data access beyond the usual suspects (research institutes, pharma, biotech companies) to harness untapped resources—such as toxicity and trial data that pharma companies have been reluctant to share as well as hospital data that includes real patient interactions. With sensible defenses built in to ensure compliance and protect privacy, accessing complex, real-world outcomes from hospital data can open up new horizons for researchers seeking to create better treatments.
Craig Monell, PhD, Senior Vice President, Business Operations, BioLegend (Part of Revvity)
Cellular multiomics advances shaping life sciences and diagnostics: The coming year promises to be a transformative one for the life sciences, with cellular multiomics taking a leading role in helping to advance our understanding of cellular composition and function, and potentially revolutionizing clinical diagnostics. In the rapidly evolving field of cellular multiomics, we can anticipate significant breakthroughs that will shape the future of life sciences. Cellular profiling assays are transitioning from traditional reliance on a limited set of markers within a singular molecular class to newer high-plex, high-cell number multiomics assays. These assays target a combination of RNA, proteins, DNA, and various other marker types, and they are deepening our understanding of cellular composition and function. We can expect this very active applications development area to deliver significant advances in assays including expanded marker numbers and diversity; increased numbers of cells interrogation; streamlined workflows; enhanced resolution in spatial analysis; and reductions in the cost per cell or marker analyzed. These improvements will make multiomics technologies more accessible and drive their adoption. Additionally, AI/ML-driven informatics will simplify the interpretation of complex multiomics datasets, making it easier to extract meaningful insights from vast amounts of data. While researchers in the highly competitive life sciences arena are rapidly making progress, we can also expect to see CROs and life science tool providers working diligently to standardize specific assays. This standardization is a crucial step toward incorporating these technologies into clinical trials and eventually deploying multiomic signature-based diagnostics and companion diagnostics.
Niven R. Narain, PhD, President and CEO, BPGBio
In 2024, we anticipate entering a new phase in the realm of AI-driven drug discovery, marking the onset of a post-hype era: While the term “AI in drug discovery” has become ubiquitous, the focus is shifting toward a more discerning evaluation of success—measured not by mere adoption but by tangible achievements in drug development through AI. The prevailing narrative has seen widespread integration of AI in drug discovery, yet the crucial distinction lies in those who have effectively harnessed this technology to guide and contribute to the discovery of novel drugs. As the industry matures, the emphasis is shifting from the mere use of AI to the substantive impact it has in propelling the discovery and development of groundbreaking pharmaceuticals and diagnostics, which is a necessary and welcome change.
Christopher Bouton, PhD, SVP and Head of AI, Certara
AI will get a biomedical PhD: Drug developers will realize that they’re capturing just a fraction of the value sitting inside their most hard-won asset: their data. They’ll race to utilize proprietary GPTs, trained on both their data and publicly available data, to surface insights that propel every stage of development, from compound identification to clinical study report authoring.
Peter Ellman, President and CEO, Certis Oncology Solutions
Virtually every pharmaceutical company is now using AI to identify new biomarkers, find new targets, design novel compounds, or look for new uses for existing medicines: But as AI has taken center stage in corporate narratives, a new challenge has emerged, namely, validating these predictions. As we see continued adoption of AI in drug discovery in 2024, we also expect to see an increase in demand for pharmacology studies using more clinically relevant animal models such as orthotopic patient-derived xenografts (O-PDX), and humanized O-PDX models will instill confidence in AI predictions and accelerate machine learning.
Alister Campbell, Vice President Global Head of Science, Dotmatics
Pharma will keep closing the digital transformation gap: Life science companies will start to see the fruits of digitalization as they increase the use of drug repurposing, making it easier to navigate the swaths of data that have previously been locked away in silos, unlocking the potential of drugs that have been shown to be safe but missed initial clinical endpoints. Healthcare companies have increased their digital capabilities more than any industry except consumer goods since 2019. Under pressure to deliver new treatments and to be multi-modal in their research, life sciences companies will continue to break down silos that impede data sharing and digital collaboration. Pharma companies did the hard work of establishing a shared vision for digital — now they need to make that vision a reality so they can reap the benefits of AI. They need to push each other and the vendors that support them toward use of technologies that ultimately support multi-modal research driving toward an AI future.
IP becomes more critical than AI: Over time, AI in discovery will become table stakes. In its current incarnation it will move the needle only slightly in early optimization but as yet it is not addressing the challenges in later stage development and clinical studies. Intellectual property is a pharmaceutical company’s lifeblood. It represents the company’s unique knowledge, gained through decades of tireless work by researchers and scientists. The increasingly digital nature of the drug discovery process means that IP is now encoded in data. Conversely, AI is a commodity. Vendors are building exciting AI products on top of public knowledge. Their models make it easier to glean relevant insights from chemical structures and properties of molecules. These products push the starting point for research forward for every company that buys them. AI only becomes a unique advantage when it’s deployed atop a foundation of proprietary data that is optimized for use in R&D. Optimization, in this context, refers to both technology and governance. AI needs clean, standardized, organized data to produce usable outputs. In turn, companies need a strong data governance culture to ensure that the data produced and collected by different teams are consistent. But when computational models are used to design new drugs, the debates will ensue around who owns the IP if it was derived from AI?
We will begin to see the long-awaited move toward shorter and less expensive drug cycles: The discovery of the COVID-19 vaccines showed what is possible when organizations work together, pooling resources and data to enable quick identification and development of novel therapies. In addition, the streamlining of the regulatory processes made it possible to bring lifesaving vaccines to the market. The current drug discovery process is complicated and inefficient: bringing a single drug to market costs around $2.5 billion dollars, takes 10 years, and for every drug that gets approved, ten thousand compounds will fail. Industry research has found that over the past 20 years, large pharmaceutical companies have spent an average of $6.16B per drug approval. Technology and innovation is now helping life science companies shorten the drug discovery funnel and reduce the costs of drug therapy R&D from billions to millions of dollars.”
Stephanie Franco, PhD, Senior Scientific Affairs Liaison, EUROIMMUN (Part of Revvity)
Blood based biomarkers and genetic screening will gain traction as complementary to disease modifying therapies (DMTs) for patients with early-stage Alzheimer’s Disease: With the excitement surrounding the FDA approval of Leqembi and pending approval of donanemab (anti-amyloid treatments for Alzheimer’s disease (AD)), we see further clinical investigations into blood-based biomarkers (BBBMs) and genetic screening as imperative to furthering improvements in AD diagnosis and treatments. It has been demonstrated that APOE e4 carriers are at a higher risk for developing amyloid related imaging abnormalities (ARIA) when undergoing anti-amyloid treatments, such as Leqembi and donanemab. Thus, it has been recommended in the Leqembi instructions for use to perform APOE genetic screening prior to treatment, which suggests this genetic screen could become a requirement for treatment. Furthermore, several new and existing biomarkers in plasma (e.g p-tau-217) are being investigated as having value in accurate diagnosis and treatments which are more economically friendly, easily accessible and less invasive than CSF draws and PET screens. We also see further investigations into combination therapies with targets other than amyloid to help reduce cognitive decline in patients with mild to moderate AD. Further exploration of BBBMs and APOE genetic testing are surely to be pivotal in the next chapters of AD diagnosis and treatments. The changing landscape of Alzheimer’s disease research and clinical trial data presented at the most recent Clinical Trials for Alzheimer’s Disease (CTAD) conference was described as profound and we agree.
Mark Kiel, MD, PhD, Chief Scientific Officer and Vice President of Product Strategy, Genomenon
Personalized medicine: We’ve witnessed tremendous advances in high-throughput genetic sequencing and genetic editing with CRISPR-based tools. The translation of accumulated knowledge from these technologies to safe and efficacious therapies has lagged behind, however. I predict that will change in the coming year as research tools that harness AI for scientific literature mining continue to evolve and accelerate our understanding of human genomics and rare genetic diseases. An important part of that evolution will be recognizing the limits of AI and devising solutions that surmount them!
Ariel Katz, CEO & Co-Founder, H1
The industry will experience a reckoning that AI is here to supercharge healthcare professionals and augment their work, not replace them: Pharmaceutical companies are on the brink of a tech revolution, and will embrace AI as a means to enhance their operations and research capabilities. In the near term, AI will be all about saving time, simplifying work processes, and knocking down language and jargon barriers in healthcare and clinical research. Imagine a healthcare world where medical research is crystal clear and open to all, making science feel like a breeze. This is the power of AI.
Surya Singh, MD, CEO of InformedDNA
Interpretation of -omic tests already being performed will improve, driven by better analytics and use of human expertise: Genetic test results must be interpreted in a way that is: a) correct, b) usable for community physicians, and c) easily explained to patients. While the potential for using AI-enabled tools to supplement human interpretation is large, these tools simply can’t deliver the same understanding, empathy, and integrative path forward that skilled clinicians—such as genetic counselors and pharmacists—can offer. Therefore, person-to-person communication and conversations will remain crucial to the success of precision health in oncology, pharmacogenomics, maternity care, and beyond.
Jesse Mendelsohn, Senior Vice President, Model N
AI will supercharge miracle drug development: “In the next few years, we will see a surge in highly effective, multipurpose drugs. Pharmaceutical companies will leverage AI to supercharge drug design with greater precision, specificity, and speed, resulting in more effective treatments with fewer side effects. Very soon, more new medications will move from treating the symptoms to fixing the root problem.”
Kimberly Powell, Vice President of Healthcare, NVIDIA
Generative AI drug discovery factories: A new drug discovery process is emerging where generative AI molecule generation, property prediction and complex modeling will drive an intelligent lab-in-the-loop flywheel, shortening the time and improving the quality of clinically viable drug candidates. These AI drug discovery factories employ massive healthcare datasets use whole genomes, atomic resolution instruments and robotic lab automation capable of running 24/7. For the first time, computers can learn patterns and relationships within enormous and complex datasets and generate, predict, and model complex relationships in biology that for the drug discovery industry was only possible through experimental observation and human synthesis.
Paul Hawkins, PhD, Principal Product Evangelist, OpenEye, Cadence Molecular Sciences
In 2024, we foresee predictive modeling methods falling into two classes, physics-based methods that calculate the property of interest of a given molecule, and machine leaning or AI methods that recognize a molecule that possess that property: While AI has been successfully applied to a wide range of problems, success in small molecule drug discovery has been (relatively) slow and patchy. The recent failures of several ‘AI-developed’ drugs in clinical trials serves as a reminder that drug discovery is difficult and complex. A particular hurdle in the application of AI to drug discovery is the paucity of data. Even in highly resourced projects, only a few thousand data points will be gathered.
Several avenues to improve AI’s performance in drug discovery will continue to be explored. One approach is to reduce data hunger by using AI methods that require only a small number of examples (few-shot learning) or a single example (one-shot learning) with a desired property. Another pathway is to use physics and AI to augment one another. Physics-based methods can provide complex featurization, such as free energy or electrostatic potential. Using physics to calculate molecular properties at a scale inaccessible to experiment are then used to train more data-intensive AI models.
The drug discovery community will better understand that AI methods themselves can also be thought of as one of two categories, transparent or opaque, based on human ability to understand the basis of the predictions a method makes. Transparent methods, such as random forest (RF), permit the user, at least in theory, to follow the chain of reasoning that led to a prediction and possibly to learn from it. Opaque methods, such as large language models (LLMs), do not permit human interpretation and learning and therefore must be used as a black box. We predict an improvement in the interpretability of models, particularly through graphical explainers, will substantially improve the adoption and the impact of AI across drug discovery.
Chris Learn, PhD, PMP, Senior Vice President, Cell and Gene Therapy, Paraxel
In 2024, I anticipated continued growth of cell and gene therapy clinical development: There are numerous factors driving growth in cell and gene therapy clinical development. First, there are an increasing number of therapy assets, with approximately 1,200 selling gene therapy assets in clinical trials and another 2,800 potentially coming into the clinic within the next 2 to 5 years. This influx of therapies creates opportunities for learning, growth and development in the field. We are also seeing more suppliers entering the market and greater market entry, so it is expected that the cost of these therapies will go down. With cost reductions, access and commercialization opportunities will improve. Finally, because cell and gene therapies have shown meaningful clinical outcomes in monogenic disease states, we have seen the possibility for these therapies to greatly improve quality of life and reduce healthcare costs for large patient populations with chronic diseases in addition to terminal ones. This, too, is driving further growth and opportunity.
Michelle Fraser, Head of Cell & Gene Therapy, Revvity
Gene editing & CRISPR technologies propel cell and gene therapies into a new era of precision and complexity: In the dynamic realm of gene editing and CRISPR technologies, the pace of advancement is nothing short of remarkable. Developers in the cell and gene therapy arena are fervently working to create improved therapies characterized by greater efficacy, improved safety profiles, and streamlined manufacturing processes. At the heart of this pursuit lies innovations in payload and delivery mechanisms. Within the gene editing landscape, developers are narrowing in on novel nucleases with specificity to, and optimal activity, at the targeted site. This endeavor seeks to navigate regions that may elude the original Cas9 nuclease. Simultaneously, developers are honing deaminases with high editing efficiency, coupled with a meticulous avoidance of bystander or off-target editing. The quest for precision extends to the realm of guide RNA design algorithms, ensuring the accurate navigation of nucleases to their intended target sites.
The evolving toolkit is empowering therapeutic developers to engineer improved therapies. This includes the capability to knockout multiple genes, suppress or upregulate specific genes, or insert new copies of genes. Notably, there are also technologies poised to edit mRNA rather than the genome, inducing transient changes, and addressing genetic alterations influenced by environmental factors through epigenetic modifications. As our comprehension deepens regarding the intricate interplay between the genome, proteins, and cellular functions, coupled with the expanding arsenal of genome and transcriptome manipulation tools, the landscape of cell and gene therapies is witnessing a paradigm shift. The trajectory points towards increased sophistication, efficacy, and safety in these therapeutic interventions.
Anticipating the trajectory in 2024, the horizon is set for substantial strides in gene modulation technologies. Expect to see a surge in the adoption of more intricate gene editing approaches, signaling a year marked by significant advancements in the refinement and application of these transformative technologies.
Cell and gene therapy approvals surge: In the rapidly evolving landscape of cell and gene therapies, the momentum is palpable. According to ASGCT, the pipeline boasts more than 3,700 therapies in development, setting the stage for a transformative period. Forecasts from TuftsMedicine further underscore this anticipation, highlighting the likelihood of at least five FDA approvals by the close of 2023, with a notable emphasis on the promising realm of CAR-T/TCR therapies for blood cancers. The statistics reveal a compelling narrative – CAR-T/TCR therapies exhibit a more than twofold likelihood of approval upon entering Phase I compared to their hematological oncology counterparts. Similarly, orphan gene therapies demonstrate double the chances of approval post-Phase I trials, outpacing average drugs in analogous therapeutic areas.
A noteworthy acceleration is evident, with over 500 new cell and gene therapies entering clinical trials annually. Their expedited time to market, averaging five years as opposed to the 10+ years for traditional small molecule drugs, positions this class of therapeutics for significant prominence in the pharmaceutical landscape. As we approach 2024, the prospect of witnessing a surge in approvals becomes increasingly imminent. Concurrently, the FDA’s proactive initiatives to bolster the advancement of these groundbreaking therapies promise to usher in a new era of standardization. Expectations include the streamlining of manufacturing processes, consolidation of protocols for testing Critical Quality Attributes, refined approaches to patient identification and enrollment, and the evolution of companion diagnostics. This concerted effort towards standardization is poised to shape a more efficient and robust framework for the evaluation and deployment of novel therapies, marking a pivotal step towards their widespread adoption in the coming years.
Guy Guzner, Co-Founder and CEO, Savvy
SaaS will democratize the IT department: Similar to how social media democratized the news, SaaS is poised to democratize IT. Third-party productivity and generative AI offerings are acting as gateways for employees to expect the freedom to leverage any tool to get their work done, irrespective of IT policies. This will create significant organizational challenges as IT and security teams grapple with the mass proliferation of unsanctioned SaaS usage. Ensuring that basic identity hygiene is maintained, identifying the lack of SSO and reused passwords, as well as effectively off-boarding users when they leave will become both more difficult and more important than ever.
CISOs will begin to franchise cybersecurity: We are starting to see CIOs franchise digital delivery and co-lead, co-produce and co-deliver digital transformation initiatives. This evolving mindset will extend into the realm of security, where CISOs are poised to take on a more collaborative and co-owning role with CIOs and business unit leaders in the development and decision-making of security solutions. This departure from the sometimes adversarial relationship between CIOs and CISOs is indicative of a broader trend toward integrated governance.
Todd Dickinson, PhD, CEO, Stellaromics
In 2024, the field of spatial biology is on the cusp of a paradigm shift, with advancements set to reshape biomedical research: To date, spatial technologies have been limited to two-dimensional data taken from extremely thin slices of tissue, resulting in significant loss of valuable information. Moreover, the platforms of today are largely limited to generating static gene expression measurements, rather than probing downstream functional readouts such as which genes are actually being translated to proteins. In 2024, I believe we will see the advent of true 3-dimensional, functional spatial technologies emerge. Technologies like STARmap and RIBOmap, which deciphers 3D transcription and translation patterns in thick native biological tissue, will help drive this evolution in the field.
Jo Varshney, CEO and Founder, VeriSIM Life
Half of new drug applications will feature AI, but translation will remain a challenge: In 2023, the FDA issued a request for industry comment on the integration of AI in drug and biologics development. They acknowledged the growing use of AI by sponsors, noting that more than 100 IND submissions received in 2021 included AI/ML components. We believe that 2024 will see an increase to ~400 applications referencing AI/ML data and analyses. But despite the increased prevalence, AI-enhanced drug candidates may not fare well in the clinic, as was evidenced in 2023. This is because much of the early application of AI has been on molecule discovery. And while using AI to explore chemical space and unlock new targets is valuable, its potential to guide translational research has been far less adopted. Sadly, this bottleneck still prevents the accelerated access to novel therapies for patients in need, as clinical trials fail from endpoint misses and toxicity. In 2024, pharma companies and emerging biotechs will improve candidate safety profiles and reduce off-target effects by exploiting predictive intelligence unlocked by AI.
