The AI Lab Rat: How Algorithms Are Becoming Life Science’s Newest Guinea Pigs

4/11/24

Editorial team at Bits with Brains

The integration of artificial intelligence (AI) into drug discovery is a paradigm shift in how life science companies approach the development of new medicines.

Traditionally, the process of developing new therapeutics has been labor-intensive, time-consuming, and fraught with high failure rates. AI, particularly machine learning and deep learning algorithms, provide a mechanism for streamlining this process by analyzing vast datasets to identify potential therapeutic candidates more efficiently than before. This not only accelerates the pace of discovery but also has the potential to significantly reduce costs and increase the success rate of new drugs entering the market.

Generative AI models, the current darlings of artificial intelligence, are particularly intriguing for their ability to create new data instances that resemble the training data. In the context of drug discovery, these models can design novel protein structures that could serve as potential therapeutics. This capability is groundbreaking because it moves beyond the mere analysis of existing biological compounds to the creation of entirely new ones. Such advancements could lead to the discovery of drugs with novel mechanisms of action, addressing diseases that currently have limited treatment options.

A good example is Profluent. The company uses AI to move from accidental discovery to intentional design of medicines, focusing on genetic diseases that cannot be addressed by naturally occurring proteins or enzymes. Profluent's technology is validated by a peer-reviewed paper in Nature Biotechnology, demonstrating one of the first uses of large language models (LLMs) to generate entire proteins that function in real-world applications.

Collaboration between AI research institutions and life science companies is crucial as well for translating AI's theoretical potential into practical therapeutic solutions. Such partnerships leverage the computational and analytical prowess of AI with the biological expertise and experimental capabilities of biotech firms. By combining these strengths, the drug discovery process becomes more efficient, from initial target identification to preclinical testing. This collaborative approach not only accelerates the development of new drugs but also ensures that the AI models are grounded in biological reality, enhancing their relevance and applicability.

The development and application of AI in drug discovery has been significantly bolstered by investments and support from industry leaders and institutional investors. Financial backing is essential for fueling research, developing sophisticated AI models, and conducting necessary experiments. Further, endorsements from prominent figures in the tech and AI fields lend credibility and attract further attention and resources to these initiatives. This ecosystem of support is vital for overcoming the substantial technical and financial challenges inherent in drug discovery.

As AI technologies continue to evolve, their application will likely broaden to include personalized medicine, where treatments are tailored to the individual genetic makeup of patients, and the discovery of multi-target drugs that can address complex diseases like cancer more effectively. Furthermore, AI could play a role in predicting drug efficacy and side effects, improving clinical trial design, and even in the manufacturing process, ensuring drugs are produced more efficiently and safely.

For AI-forward life science organizations, the implications of implementing AI in drug discovery will be profound. The first step for these organizations is to rapidly and critically evaluate the specific AI technologies that are most applicable to their research and development needs, such as machine learning algorithms and models capable of processing complex biological data.

Secondly, AI-forward organizations should focus on developing robust data infrastructures and ensuring access to high-quality, diverse datasets. AI models are only as good as the data they are trained on, so executives must prioritize data governance and quality. This includes establishing partnerships with healthcare providers, research institutions, and possibly even competitors to enrich their data pools. Additionally, they must focus on addressing data privacy and security concerns, ensuring compliance with regulations such as HIPAA in the United States or GDPR in Europe.

Third, these organizations should accelerate a culture of continuous learning and adaptability. AI is rapidly evolving, and staying abreast of the latest developments is challenging, but crucial. This may involve setting up dedicated AI research teams and investing in ongoing training and reskilling of their staff. By doing so, they can maintain a competitive edge and quickly adapt to new AI advancements that could further enhance drug discovery efforts. Rapid pilot projects that either succeed, or fail quickly, are key.

Conversely, for life science organizations slower to adopt AI, the challenges are likely to become more pronounced with time. In some industries, early AI-adoption is a winner-take-all proposition, and executives need to understand and internalize this.

The first step for executives in these organizations is to recognize the strategic importance of AI in maintaining competitiveness in the life science industry. They must critically but rapidly assess the potential return on investment that AI offers in terms of faster drug discovery timelines and reduced R&D costs. Time is not their friend.

The second step is to immediately begin building the necessary infrastructure and expertise required to implement AI effectively. This may involve hiring new talent with AI expertise, upskilling existing employees, seeking external partnerships. Executives must also be prepared to make the necessary capital investments in technology and data management systems. They too should start with pilot projects to demonstrate the value of AI in drug discovery, which can help build internal support and momentum for broader AI initiatives. And all of this needs to be done rapidly.

For many late adopters of AI, M&A will become an essential tool, and possibly the only tool, to reduce the existential threat from their AI-forward competitors.

Sources:

[1] https://techcrunch.com/2024/03/25/profluent-spurred-by-salesforce-research-and-backed-by-jeff-dean-uses-ai-to-discover-medicines/

[2] https://www.politico.eu/article/ai-is-about-to-remake-the-pharmaceutical-drug-medicines-industry/

[3] https://www.knowmadmood.com/en/blog/12-uses-of-artificial-intelligence-in-the-pharmaceutical-industry/

[4] https://arxiv.org/ftp/arxiv/papers/2212/2212.08104.pdf

[5] https://www.frontiersin.org/articles/10.3389/fbinf.2023.1121591/full

[6] https://www.nature.com/articles/d41586-023-03172-6

[7] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10385763/

[8] https://www.sartorius.com/en/knowledge/science-snippets/the-trending-role-of-artificial-intelligence-in-the-pharmaceutical-industry-599278

[9] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10302890/

[10] https://www.spiceworks.com/tech/artificial-intelligence/guest-article/top-challenges-faced-by-pharma-ai/amp/

[11] https://www.pharmaceutical-technology.com/data-insights/artificial-intelligence-in-pharma/