Trendspotting: What’s Coming for Bio-IT in 2023
January 5, 2023 | We spoke with several Bio-IT World community leaders to gain insights about their predictions for the coming year. For the first time in two years, COVID was not the primary topic of conversation. Instead, using artificial intelligence (AI) and large language models (LLMs) to accelerate workflows and develop novel therapies are expected to be the primary focal points this year. “The hype about generative AI becomes a reality in 2023. The foundations for true generative AI are finally in place, with software that can transform large language models and recommender systems into production applications that go beyond images to intelligently answer questions, create content, and even spark discoveries. This new creative era will fuel massive advances in personalized customer service, drive new business models and pave the way for breakthroughs in healthcare,” said Manuvir Das of NVIDIA.
AI will also play a more significant role in how big data is curated and managed. Miruna Sasu of COTA Healthcare said, “This industry will make significant advancements in using technology to scale abstraction and data curation—from electronic health record data and claims data to new types of data like genomics and patient-reported outcomes. As we reach the point of accessing and deriving insights from big data in health and medicine, AI will continue to get smarter and more impactful for patients and industry alike.”
Most biotech leaders agree that AI, computing, and human curation will be necessary to remain successful in the industry. Charlie Boyle of NVIDIA said, “Amidst economic headwinds, enterprises will seek AI solutions to deliver on objectives while streamlining IT costs and boosting efficiency. New platforms that use software to integrate workflows across infrastructure will deliver computing performance breakthroughs—with a lower total cost of ownership, reduced carbon footprint, and faster return on investment.” Mark Kiel of Genomenon emphasizes the importance of striking a balance between automated and human-led processes: “Rather than move away from the need for human curation, the tension between optimized sensitivity and specificity will promote a flourishing of novel curation capabilities to augment human review, leading to an effective merging of AI computation and human curation.”
There will also be a concerted effort to conserve resources and maximize efficiency. “This year, software will enable businesses to unify AI pipelines across all infrastructure types and deliver a single, connected experience for AI practitioners. This will allow enterprises to balance costs against strategic objectives, regardless of project size or complexity, and provide access to virtually unlimited capacity for flexible development,” said Das.
Of course, the industry will continue making strides in whole-genome studies. Christian Henry of PacBio said, “High-accuracy long-read sequencing has finally enabled comprehensive studies of genomic variation across the entire human genome. This further opens the door to genuinely diverse projects touting pangenome analysis from previously under-represented populations, including the full range of genomic diversity across the globe that challenge developers to create tools to capture these discoveries.”
Here are the full trends and predictions, including additional forecasts for joining point solutions in digital health, high-accuracy long-read sequencing, regulatory initiatives for data management approaches, and a shift toward lightweight algorithms that do not require massive amounts of data. –the Editors
Christian Henry, President and Chief Executive Officer, PacBio
Reference-quality human genomes for all: The post-telomere to telomere era will commence with researchers being able to benchmark DNA sequence data against the entire human genome (Science, DOI: 10.1126/science.abj6987). HiFi long-read sequencing will enable researchers to interrogate these new genomic regions—including dark regions—in both existing and large-scale studies and potentially identify more medically relevant genes. High-accuracy long-read sequencing has finally removed this technological barrier, enabling comprehensive studies of genomic variation across the entire human genome. This further opens the door to genuinely diverse projects touting pangenome analysis from previously under-represented populations, including the full range of genomic diversity across the globe that challenge developers to create tools to capture these discoveries.
Miruna Sasu, President and CEO, COTA
We’ll continue to see more acquisitions and consolidation in digital health funding: This year will be the “Year of Consolidation.” Investors will look for smart ways to join point solutions in digital health—almost like a string of pearls. These investors are eager to integrate proven digital health point solutions that, under one roof, can prove scalable and impactful in addressing serious healthcare challenges, such as chronic diseases like cancer.
There will be a major shift in AI: This industry will make significant advancements in using technology to scale abstraction and data curation—from electronic health record data and claims data to new types of data like genomics and patient-reported outcomes. As we reach the point of accessing and deriving insights from big data in health and medicine, AI will continue to get smarter and more impactful for patients and industry alike.
Jim Reilly, Vice President of Development Cloud Strategy, Veeva Systems
Simplicity and technology intersect to streamline drug development: Operational simplification and technology advancement will create cross-functional efficiency across clinical, regulatory, quality, and safety. This will enable biopharma to create a more streamlined drug development process rooted in lean process execution and higher-quality data. Connected data across the development lifecycle will enable different functions to coordinate decisions, and a common technology framework will eliminate duplicate data capture and inefficient processes. In addition, automated workflows, data reuse, common training, and a simpler technology experience will help companies adapt quickly to changing market conditions and deliver products more efficiently.
Manuvir Das, Senior Vice President of Enterprise Computing, NVIDIA
Software advances end AI silos: Enterprises have long had to choose between cloud computing and hybrid architectures for AI research and development. Unfortunately, this practice can stifle developer productivity and slow innovation. This year, software will enable businesses to unify AI pipelines across all infrastructure types and deliver a single, connected experience for AI practitioners. This will allow enterprises to balance costs against strategic objectives, regardless of project size or complexity, and provide access to virtually unlimited capacity for flexible development.
Kimberly Powell, Vice President of Healthcare, NVIDIA
Biology becomes information science: Breakthroughs in LLMs and the fortunate ability to describe biology in a sequence of characters allow researchers to train a new class of AI models for chemistry and biology. The capabilities of these new AI models give drug discovery teams the ability to generate, represent, and predict the properties and interactions of molecules and proteins—all in silicon. This will accelerate our ability to explore potential therapies' infinite space.
Kevin J. Knopp, CEO and Co-founder, 908 Devices
Automation, AI, and machine learning (ML) will revolutionize biopharma: Today, it takes 10 years for a drug to get to market, but the advent of AI and machine learning has the potential to accelerate workflows, thereby transforming biopharma development and manufacturing. Connected online devices that stream process analytics empower users with more real-time quantitative and qualitative assessments. With the advancement of personalized medicine, manufacturers will leverage technology to meet the need for greater specificity and simultaneously improve their workflow and analytics capabilities. Automated devices with embedded machine learning/AI that are simple to use will provide robust analytics and drive significant quality, efficiency, and cost improvements to the biopharma industry as well as better drug discovery and delivery to patients.
Carola Schmidt, General Manager of Automated Robotic Solutions, PerkinElmer
Achieving operational excellence through automation: Considering the issues associated with today’s laboratory personnel shortage, labs continue to be challenged by the ever-changing quality and regulatory requirements. Fortunately, technological advances have paved the way to combat these obstacles in the form of automation. As a result, we will see more automated solutions emerge, enabling labs to free up personnel for more sophisticated work where their expertise is needed. These solutions will also improve the accuracy and throughput required to make critical, time-sensitive clinical decisions 24/7. Therefore, it will be essential to seek out SMART (scalable, modular, agile, reliable, and tailor-made) lab automation to accommodate various needs.
Embracing data-driven approaches to supplement manual processes: As AI/ML-based technologies evolve, more companies will embrace data-driven approaches to replace or supplement existing manual research and development procedures of novel therapeutics for patients who have breast cancer, amyotrophic lateral sclerosis, and other similarly structured genetic diseases. Capitalizing on the recent success that the RNA-based vaccine had against COVID, we will see that the biopharma industry will implement AI/ML-based technologies across all research and development areas exponentially. Large deals featuring AI/ML biotechs will continue to close. In 2023, we will see how AI/ML-identified targets and drugs will be clinically tested and validated. The market will take note.
Next-generation sequencing: The emergence of AI/ML in biopharma and the growing availability of next-generation sequencing datasets have allowed significant bandwidth for novel targets and drug discovery. In addition, the ability of AI/ML to process large volumes of datasets enables drug development to become more information precise and risk-averse which translates to higher probabilities of success in downstream validations and development stages toward eventual regulatory approval.
Drug discovery and development: The biggest breakthrough will be when we have enough data from many different sources (i.e., molecular data, genetic data, patient data in electronic health records, claims data, clinical trial data, real-world data, etc.). Only then will we be able to see real breakthroughs at scale fully. We will have many wins along the way, but the biggest breakthrough will be getting all this data together consistently and efficiently.
Effective resource management: In a study published in 2021 on global healthcare resource efficiency, the authors concluded that healthcare systems around the globe needed to “throw resources” at the pandemic, likely raising inefficiency through wasted resource use (Frontiers in Public Health, DOI: 10.3389/fpubh.2021.638481). However, AI-based systems have been successfully used to manage hospital resources. Take Saint Luke's Health System in the Kansas City region, where the use of an AI-enhanced platform allowed the organization's flagship hospital to accommodate 7% more surgical cases despite having 20% fewer operating rooms due to staffing challenges.
Pradeep Ramesh, Principal Machine Learning Scientist, Sherlock Biosciences
AI will become more data efficient in the life sciences: Instead of asking, “How do we get more data to inform our algorithms?” I expect to see more people in 2023 asking, “How do we build algorithms that don't require massive volumes of data in the first place?” This pivot towards more lightweight—yet equivalently predictive models—accelerates the decentralization revolution by enabling researchers who do not have access to supercomputers to tinker and explore new ideas in ML, thereby accelerating the pace of scientific progress. Rather than hunger to increase the parameters in an algorithm, there will be a focus on distillation and reducing AI’s complexity to its bare-bone essence by stripping out redundancy and waste. Consequently, we can leverage trained models to generate testable hypotheses, advancing fundamental insight. The question then becomes where we need data and not simply how much.
Sivan Bercovici, CTO, Karius
Biggest trends in AI: We will see and experience both the added value and the negative ramifications of lowering friction, eventually giving rise to more robust quality control data and ML platforms. ML tech will trickle down to all corners of life, and refinements to our regulatory understanding will be shaped by the delicate balance between the risks and benefits as we allow the machines to take more on. We will get even more co-pilots across our personal and professional spaces. The biggest opportunity is in coordinated AI agents that go beyond being sophisticated alerting systems, connecting the dots across AI machines and data repositories, and taking action. Naturally, the expansion into the "action" space will occur gradually, but it is an opportunity that is bound to be explored. We all deserve access to the best care, and expanding into action will be an essential step toward that future.
Anand Parikh, Co-Founder and CEO, Faeth Therapeutics
The funk is here to stay (for a few years): The biotech industry has not stepped out of its funk. Public stocks have recovered (a bit), but private valuations are still rough, and the IPO market is closed shut. At the beginning of 2022, people believed that one large acquisition or IPO would help snap the market out of its funk. Biotech investment is risky, and when interest rates are low, that risk is much more attractive. However, when short-term government bonds pay 5% interest rates, the risk/return of investing in biotech can look less palatable. Therefore, big pharma is less likely to snap biotech out of the doldrums; instead, the industry will be more dependent on broader consumer sentiment improving and the federal reserve beginning to lower interest rates. It might be towards the beginning of 2024 before we start to see interest rates decrease.
Paul Steinberg, Chief Commercial Officer, Resolve Biosciences
A wave of new personalized therapies: The ability to analyze individual cells and their interactions within tissue in a three-dimensional spatial context with high sensitivity and subcellular resolution will open the door to a wave of scientific findings that will redefine our understanding of disease and lead to new personalized therapies and improved patient outcomes.
Sara Patterson, Ph.D., Associate Director of Clinical Genome Informatics Products, The Jackson Laboratory
Precision immuno-oncology: We expect to see a new wave of precision immuno-oncology therapies due to new tools and technologies that improve our ability to collect clinically actionable data and rapidly analyze complex tumor profiles.
Emmanuel Abate, President of Genomic Medicine, Cytiva
Genomic medicine is an area where we will continue to see increased investment in technologies and new clinical trials: From cell therapies to gene therapies to mRNA vaccines, genomic medicine has the potential to transform global health. One of the most exciting aspects of genomic medicine is that it can treat the causes of disease rather than just symptoms. However, to realize its full potential, the industry must address the current manufacturing and supply chain challenges. We must also look at how we can better deploy digital solutions to accelerate the manufacture of these novel therapeutics. 2023 will be an exciting year as the biotechnology industry works to solve some of the world’s most complex health challenges.
Alan Jiang, CSO, XtalPi
Automation and AI: a marriage made in disruptive technology heaven: In the past decade, no other technology has garnered more attention than AI. AI is an indispensable infrastructure, accelerating and amplifying computational efforts. Automation also needs no introduction. Across the world, automation has taken over repetitive tasks in manufacturing, saving both time and money. But the real magic happens when we combine automation with AI. What better source of accurate, standardized, error-free, and reproducible data than automated robots that do exactly what they’re told, twenty-four-seven, without the need for rest or vacation? With AI trained on sufficient, homogeneous data, we can accurately anticipate synthetic outcomes and eliminate wasted time and resources on unproductive attempts. Once fully trained, synthesizability can become another screening parameter in the early stages of lead discovery, further enhancing the success of delivering promising drug candidates.
Alan Kalton, Global SVP, Aktana
The chasm will widen between those companies investing in analytics-based, digital-first commercial models and those still sitting on the fence and taking a measured approach: COVID-19 put many companies in the uncomfortable position of forcing new ways of working, engaging with customers, and new technologies. During the pandemic, we saw the impact of adopting new technologies, such as platform intelligence solutions “on the ground,” and 89% of companies surveyed by DHC Group reported successfully executing an AI-driven omnichannel strategy across sales and marketing and scaling up. Today, many companies understand the impact of AI and recognize that it needs to power engagement across all markets. The organizations delaying investment in scaling intelligence platforms across the organization will see a widening gap between themselves and their competitors regarding influence, customer engagement, and financial success.
David Ehrlich, CEO and Chairman, Aktana
Companies will connect data science models to day-to-day operational activities to execute strategic business goals across organizations: This year, we will see the start of the next chapter in AI for commercial life sciences organizations. Traditionally, AI has lived in one of two places—either with headquarters (HQ) teams analyzing massive amounts of data to generate “smart” conclusions or within discrete applications, helping to tune the application’s impact (i.e., marketing automation systems). What’s missing is the connective tissue between HQ’s broad-scope AI to the various operational systems required to execute HQ’s strategic business goals. Operating teams could agilely deploy data science models to guide multiple day-to-day activities. Ultimately, companies will be more effective and efficient while cycling through the “try it, fix it” rhythm much faster to improve AI’s outputs across the entire organization continuously.