AI in Pharma: Navigating New Frontiers in Drug Discovery and Development

AI in Pharma: Navigating New Frontiers in Drug Discovery and Development

Bringing a new drug to market typically takes a decade or more and approximately US$2.5 billion. However, scientists are now utilising new technology, including AI, to dramatically shorten the drug discovery process and to reduce the costs of drug therapy research and development. To explore this revolutionary transformation, Pharma IQ has interviewed Alister Campbell, Vice President Global Head of Science and Technology at Dotmatics, who offers profound insights into how AI is reshaping the landscape of pharmaceutical research, making the pursuit of faster, more effective treatments a tangible reality.

What are the most significant trends in the adoption of AI within the pharmaceutical industry and how are they transforming drug discovery?

The Pharmaceutical industry is continuing to close the digital transformation gap and has done much of the hard work of establishing a shared vision for AI across the key stakeholders in the scientific community. For example, life science companies can now more easily navigate the large amounts of data that have previously been locked away in silos, which allows the potential of drugs that are safe but missed initial critical endpoints to be unlocked. However, as vast as data stores are today, many organisations are reporting that they anticipate doubling the data that they have stored in the coming year. This further necessitates the need to find ways to flexibly aggregate all relevant data into intelligent data structures enabling clean, reliable data analysis and paving the way for meta-analysis and AI & ML-based algorithms to be effectively utilised. 2024 is expected to be the year that this happens in a significant way and scientists will be further empowered to answer their most difficult and pressing questions.

With the increasing importance of big data in drug development, what strategies are essential for the industry to effectively manage and utilise this data for AI applications?  

Data generated in research organizations is by its very nature the epitome of Big Data - High Volume, High Variety, High Velocity – the 3 V’s. To manage and handle the "3 V's", scientists are required to resort to complex mathematical and statistical models to aggregate, normalise, and create insights (information) from this data. New science techniques and modalities rarely appear with a well-designed data management strategy alongside them.  Collaboration between scientists and organisations like Dotmatics is crucial in implementing FAIR data principles to accommodate new science techniques within existing data ecosystems.

How is the industry balancing speed with safety and efficacy in AI-driven drug development?

It is critical that as an industry, we convene and address honest, actionable discussions about the challenges and the safeguards of AI use in drug discovery. Ensuring that scientists using AI have reliable underlying data and ensuring that the data used to train AI is clean, organised, and unbiased, is paramount. AI is only as good as the information that it learns from, and bad data leads to bad science. Additionally, there will be a growing emphasis on scientists to capture where AI/ML models were used and the training sets that were used at the time.  This will enable a more accurate level of reflection on the success of the models in the future.   If a model is uninterpretable by an impartial and trained scientist, the trust in the model will be further reduced.

What steps are being taken in the pharmaceutical industry towards integrating AI, and why are incremental advances crucial?

Implementing AI in drug discovery requires many things but three stand-out steps come to mind. The first is having reasonable expectations. AI can be a valuable part of a company’s larger drug discovery program but clarifying when, why, and how AI is used is crucial, albeit challenging. Companies developing small molecules have largely received investment because they’re relatively simple compared to biologics, and because there are decades of data upon which to build models. There is also great variance in the ease of applying AI across discovery, with models for early screening and physical-property prediction seemingly easier to implement than those for target prediction and toxicity assessment. The acknowledgment that AI reaching its peak in drug discovery will likely take years is important to remind us to not rush the process but take the time needed to achieve successful use of it in future drug discovery.

The second is collecting clean data. AI requires good data, and lots of it, but good data is hard to come by. Implementing these technology and workflow processes can help in this process:

• Facilitate error-free data capture without relying on manual processing.
• Handle the volume and variety of data produced by different teams and partners.
• Ensure data integrity and standardise data for model readiness.
• Annotation of data – understanding how the data was generated will enable scientists to understand where the data can be and when it should not be compared and aggregated.

The third is collaboration - companies hoping to leverage AI need a full view of all their data, not just a part of it. This demands a research infrastructure that lets computational and experimental teams collaborate, uniting workflows and sharing data across domains and locations. Careful process and methodology standardisation is also needed to ensure that results obtained with the help of AI are repeatable.     

How will AI influence the cost of developing drug therapies?

Ultimately AI has the potential to reshape the traditional drug discovery funnel making it both wider at the top offering additional candidates and shorter in length of time to bring the right candidate to market. The combination of utilising AI together with pharma companies’ own IP can create a competitive advantage and reduce the typical 10 years and US$2.5 billion that it costs on average to bring a new therapy to the market. This efficiency could lead to more affordable drug therapies, addressing the crucial challenge of accessibility.

Reducing the time to identify clinical candidates is the first area that AI is having a significant impact.  But, once the data collected improves, the subsequent models will then improve, and we will have a much better chance of making an impact on the hardest problem - predicting success in the clinic.

What are the key regulatory challenges the industry faces with the increasing use of AI in drug discovery, and how might these be addressed?

Data integrity remains a significant regulatory challenge. The FDA's focus on data security and integrity underscores the need for technology solutions that prioritize these aspects, ensuring the safety and efficacy of drug products.

When we think of the regulatory landscape, I love this simple phrase from the NIH: “Good science requires good record-keeping". Data integrity remains a significant regulatory challenge. To ensure the safety, quality, and efficacy of a drug product before market rollout, regulatory bodies are correctly pushing the industry with regulations that will ensure that the data are complete, verified, and undistorted throughout the data lifecycle.

The FDA's increasing focus on data security and integrity underscores the need for technology solutions that prioritize these aspects, ensuring the safety and efficacy of drug products.

As the use of AI in pharma R&D becomes more mainstream, assessing data integrity will become even more complex. For biotechs to prioritise the security and integrity of their data, they need to trust and choose technology solutions that are built to inherently support these priorities as well. 

How has the pandemic accelerated or altered the trajectory of AI integration in pharmaceutical research?

The pandemic underscored the need for rapid scientific advancement and collaboration to deliver faster, more successful breakthroughs. Yet the complexity of getting and using scientific data in meaningful ways has led to compromised success and delayed results among researchers and scientists. Integrated multimodal R&D platforms with the depth, breadth, and connectivity to solve complex data challenges in labs can help to speed up access, facilitate collaboration, and inform decision-making. Dotmatics' development of platforms like Luma illustrates the industry's push towards integrated, multimodal R&D platforms to overcome data challenges and enhance research outcomes. Companies that own the cleanest, best-annotated data to power their AI analytics and decision-making will be the best positioned to succeed in this new paradigm.

What is the future outlook for AI in pharmaceutical research?

While the transformative potential of AI is immense, the integration into pharmaceutical research will be gradual. Dotmatics envisions AI significantly reducing development times and costs, empowering researchers, and ensuring responsible implementation and use of AI in scientific discovery. However, we also believe strongly that it is the IP of our customers that will ultimately make the difference in the long run when the industry is all using AI. As pioneers in scientific R&D software, we are committed to ensuring AI's responsible implementation, fostering honest discussions among stakeholders, and supporting scientists in harnessing AI's power to change scientific discovery for the better.

Looking to the Future

It is evident that the integration of Artificial Intelligence into pharmaceutical research marks a pivotal shift towards revolutionising drug discovery processes. This transformation, underscored by AI's ability to navigate the complexities of big data, enhance efficiency, and foster innovation, offers exciting potential for the development of faster, more effective treatments. Addressing the technical challenges such as data integrity, the need for clean data, and the crucial role of collaborative ecosystems, will be critical to take full advantage of the possibilities for this new era of pharmaceutical development. Moreover, keeping the principles of patient-centricity and accessibility central to any technological advancements is of the utmost importance to ensure that the maximum benefit is derived from them.

For more information on Dotmatics, visit https://www.dotmatics.com/ 

For more information on the AI revolution in the pharma industry, join the Pharma IQ community https://www.pharma-iq.com/