How Sanofi is using AI to accelerate time to market for new drugs

Artificial intelligence (AI) is transforming how scientists develop new drugs and tackle disease. Pharma IQ speaks to Matt Truppo, Global Head of Research Platforms at Sanofi, about how the pharma company is utilizing the technology to analyze large datasets and bring new drugs to market, faster.

Pharma IQ: What key things are driving the use of AI in clinical research?

Matt Truppo: Ultimately everything leads to the clinic. We need to simultaneously accelerate the drug discovery and development process while delivering higher quality and better targeted molecules with reduced attrition in the clinic, shortening timelines and lowering the overall cost overall of bringing the drug to market.

If you talk about how long it takes us to deliver a drug from concept to clinic, it is a long timeline, between three to five years.  It requires a lot of effort in terms of making chemical matter. In the case of small molecules, you might make five or 10,000 compounds before you arrive at one that you hope will be the winner. Even there, the attrition is quite high once you get into the clinic.

It involves leveraging all the data we are generating and that whole journey from concept to clinic and beyond in the most profound way. AI enables us to have insights that we couldn't otherwise gain just from using our own brain power because the datasets are just too vast.

Pharma IQ: Given that AI can help reduce the time to market, have you seen an increase in the number of drugs you are able to deliver?

MT: Yes, one area is precision medicine. When we talk about being faster, this is getting to the clinic faster and getting to patients faster. It is about aligning your resources in the right way for the best probability of success, and precision medicine allows us to do that.

When I think of precision medicine, I think of identification, credentialing and the selection of the right target. It is the design and engineering of the molecule or the medicine itself to drive better efficacy, then ultimately the selection of the patients for clinical trials and for marketing that drug – so who is going to benefit the most from that particular therapy.

When you start to think about the data in that way and how can you more precisely identify the target we should go after, that is how we are able to accelerate time to market and also gain a higher probability of success in the clinic, which ultimately leads to increased productivity.

Pharma IQ: Are there any other areas where AI has had a major influence?

MT: We are looking at how we actually embed chemical matter and the properties of that chemical matter. We have a collaboration with Exscientia, a company driven by AI and machine learning (ML). We have entered into a collaboration to focus on 15 different small molecule candidates across oncology and immunology. The goal is to accelerate the design and ultimately the analyze cycle with AI-guided molecular invention.

This is an area where we are looking at how to test an impossibly expensive structural space in the virtual world, with AI tools and models to narrow down that structural space to what we need to explore candidates for clinical trials. This has a pretty significant impact as there is no way we could make billions of compounds in the real world into smaller numbers that we can physically test to get to the ideal chemical candidate.

Pharma IQ: What predictions do you have for the next area of transformation for AI?

MT: We used to live in a small molecule world, and now we have small molecules, vaccines, proteins, peptides, gene therapies and cell therapies, so ever more complex modalities to address human disease.

Our acquisition of Amunix brought us conditionally activated biologics. For example, one clinical candidate is a masked T Cell engager that only gets unmasked in the tumour microenvironment and becomes the active drug in the presence of the tumour.

We have also acquired Synthorx which has an expanded genetic alphabet capability. They have engineered a protein that selectively targets the molecule that has non-alpha behaviour, so we can expand the right cells and T cells for oncology, while not suppressing any of their activities.

These are increasingly complex modalities. AI and ML tools have really given us an advantage to make a wide array of chemical matter, to address diseases and to target whatever mechanism of action we want.

Quick links

• Sign up to attend Pharma IQ Live: Transforming drug discovery through AI
• How to use connected innovation intelligence for drug discovery and development
• Why R&D must digitalize and become data-centric

Get exclusive access to member-only articles, reports, videos, interviews, webinars and other premium content from industry experts and thought leaders by signing up to Pharma IQ here.