AI and Drug Discovery, Part 1: Collaborations Between Pharma and AI
Part of Artificial Intelligence and Drug Discovery – A Three Part Update
A year ago, we posted about how AI was starting to be implemented in the world of drug discovery, and we predicted that AI would increasingly become integrated into the pharmaceutical industry. A year later, we can see how some of those developments are playing out: in short, AI companies are increasingly partnering with pharmaceutical companies to improve drug discovery efforts, digitize health care, and improve clinical trial development and operations. In Part 1 of this update for 2019, we outline some of the key areas of advancements and discuss some of the recent partnerships between pharma companies and AI companies. In Part 2, we will focus on an exciting and larger scale collaboration involving 10 of the largest pharmaceutical companies. In Part 3, we will identify some of the potential legal issues that are relevant to players in this space.
Increased Use of AI in Pharma
We are seeing AI starting to intersect with the pharmaceutical arena at various stages in the life cycle of drug development including drug discovery, clinical trial development and execution and post-development patient monitoring and care.
In the drug discovery front, AI is being deployed across a broad range of data sources including chemical, biological, patient data, as well as scientific literature. Companies like BenevolentAI, Atomwise, and XtalPi are focused on identifying drug leads and related target patient populations.
In terms of clinical trial design, data about patients is being used by companies including Antidote and BullFrog AI to optimize clinical processes such as recruitment, monitoring, and improving compliance of patients in the expensive clinical trial phase of drug development. For example, clinical trial data sets are being analyzed to identify which patients are likely to be most responsive to therapies and matching patients to clinical trials relevant to their conditions and characteristics. This should lead to improvements in defining inclusion/exclusion criteria, increasing enrollment, and increasing the likelihood that primary study endpoints are achieved.
Post development, AI applications are emerging in the field of digital health. Such applications include apps and software to monitor and improve patient compliance, adverse effect monitoring, and to provide patient support. For example, CardioDiagnostics focuses on wireless heart monitoring. AiCure is a smartphone app that prompts users to take their medication. Pfizer is engaging with the US Food and Drug Administration to explore AI for use in adverse-event reporting.
AI is also being implemented in patient care. For example, Novo Nordisk has a chatbot, Sofia, which uses machine learning and natural language processing to provide first-level responses to diabetes patients.
The above examples highlight the ways in which AI is poised to lead to more significant advancements in medicine and to improve overall healthcare sooner. There are various opportunities to leverage AI in the pharma industry. We anticipate that multi-pronged and diversified implementation of AI solutions throughout the life cycle of development of new drugs will be critical to remaining competitive in this new era.
Increase in AI and Pharma Collaborations
The number of public research collaborations between pharmaceutical companies and AI vendors is rapidly increasing (7 in 2016, 19 in 2017, 20 in 2018 and 30 in 2019). All of the largest 10 pharmaceutical companies (by revenue) have either partnered with or acquired AI companies.
Focus on Pharma and AI Collaborations
- Sanofi Collaboration with Google
Sanofi has entered into a collaboration with Google to establish a new virtual Innovation Lab that uses deep analytics and cloud technology to improve the drug development process. The project aims to: 1) enhance the understanding of patients and diseases by leveraging Sanofi’s trove of data and Google’s machine learning platform; 2) drive productivity gains for business operations by modernizing infrastructure; and 3) promote patient outcomes and access to healthcare through the use of technology.
- Roche Partnership with Exscientia
Roche has partnered with Exscientia in a drug discovery collaboration. Excientia’s Centaur Chemist platform uses AI to design and prioritise novel compounds for synthesis. Excientia systems learn from both existing data resources and experimental data from each design cycle. Roche retains the exclusive rights to develop and market the candidates resulting from the collaboration, while Exscientia receives upfront payments and research funding as well as commercial milestone payments.
- Bayer Partnership with Cyclica
Bayer has partnered with Cyclica to create an AI-augmented integrated network of cloud-based technologies, known as the Ligand Express. The goal of the Ligand Express is to analyze how small-molecule drugs interact with all proteins in the body. Small-molecules are screened against repositories of structurally-characterized proteins to identify potential targets and model drug effects on these targets.
- Eli Lilly and Pfizer Partnership with Atomwise
Eli Lilly has entered into a partnership with Atomwise to develop up to ten drug targets selected by Eli Lilly. Pfizer has also entered into a partnership with Atomwise for up to three target proteins selected by Pfizer. Atomwise will receive upfront payments tied to development as well as commercialization milestones. Atomwise’s structure-based deep convolutional neural network, called AtomNet is designed to predict the bioactivity of small molecules for drug discovery applications. AtomNet has already led to the discovery of two drugs that significantly reduce Ebola infectivity. These drugs were intended for unrelated illnesses and their potential to treat Ebola was previously unknown.
- AstraZeneca Partnership with BenevolentAI
AstraZeneca has entered into a partnership with BenevolentAI, hoping to combine their genomics, chemistry and clinical data sets with BenevolentAI’s target identification platform—as well as its knowledge graph systems that chart links between compounds, genes, proteins and diseases. Scientists from both companies will work together to learn more about the biological mechanisms behind chronic kidney disease and idiopathic pulmonary fibrosis.
It is clear that the opportunities for collaboration between pharma companies and AI technology companies are growing. In Part 2 of this update, we will examine a cross-competitor collaboration that recently kicked-off in Europe.
The authors would like to thank students Malcolm Woodside, Roohie Sharma, Nareesa Nathoo, and Alexandra David for their assistance in preparing this update.
 AI for Drug Discovery, Biomarker Development, and Advanced R&D Landscape Overview 2019 / Q1, Deep Knowledge Analytics, at 18.
 L.E.K. Consulting / Executive Insights, Volume XX, Issue 60 at 2 (LEK).