How PrecisionLife is using AI to improve drug discovery

Ian Bolland speaks to Dr Steve Gardner from AI-powered drug discovery player PrecisionLife, formerly Row Analytics, to talk about the company’s rebrand, its AI platform and the projects it’s involved in.  

The company recently announced a new drug discovery partnership with PharmEnable in order to develop new treatments for Amyotrophic Lateral Sclerosis (ALS) by working with King’s College London, the Motor Neurone Disease Association (MNDA) and the University of Sheffield. 

Explaining more, Dr Gardner said: “We worked with the MNDA and King’s College to get access to a genotype dataset. We ran it through our platform and we discovered a whole series of genes that were associated with ALS for the first time – 33 new genes. We then annotated all of those and we pull in data from 40 or 50 different data sources to try and prioritise them in terms of which are going to be the most druggable candidates according to the five Rs principle of drug discovery.”

The platform Gardner mentions is named after the company's rebrand. The precisionlife platform combines AI and a mathematical framework to assess how combinations of genomic, phenotypic and other clinical features relate to the development of disease at a population level. 

Scalability and speed seemingly separate PrecisionLife from other players in this space. Dr Gardner explained that their lead time spans a few months rather than three to five years – and the company has 10 targets being tested simultaneously. 

Describing the approach as “radical”, he elaborates, saying: “Every other genetic methodology that’s out there really looks at one mutation, one snip or one gene at a time. We look at combinations of genes. We know that not everybody gets a disease in the same way or progress with disease in the same way and they don’t respond to therapy in the same way. 

“Most diseases like cancer, dementia, cardiovascular disease, psychiatric disease – they involve lots of contributions from different genes, and we’re the first people at scale to be able to identify the combinations. If you can do that, you can figure out exactly what’s going on with individual groups of patients. You can test a hypothesis about what makes one group of people respond one way versus another group of people.

“What we do is look at combinations of mutations or other features and because we’re using combinations it’s giving us a lot more “buckets” to put people into. We can stratify patient populations at higher resolution, and we can then compare what’s a driving disease for one subgroup of patients versus another subgroup of patients.”

Gardner admits that PrecisionLife’s approach of requiring large amounts of data means they are unable to tackle the rarest of diseases – though largely the company’s approach allows them to be disease agnostic. 

“Big chronic diseases cost health systems a lot of money and have a big socio-economic burden plus respiratory and some of the metabolic diseases. Those are ones which we can get populations in the thousands and all of those we can analyse.

“I think this is going to be the future of medicine. We’re not going to be in a ‘one size fits all’ world, we’re going to be in a world where medicines are designed against the individual patient. You put the right drug, at the right time, at the right dose into the right patient. 

“I think if you’re going to do that you absolutely need to stratify your patients much more accurately than we can today and that’s why we set PrecisionLife up. That basic ability of putting the right things in the right patients is absolutely where we want to go.”  

Following its name change from RowAnalytics, Dr Gardner explained it was more of a reflective move from the company.

“When we started out we were still developing the underlying analytical platform and we looked a lot more like a software business. Now we are very much focussed on precision medicine as our unifying theme and that’s both the drug discovery and repurposing aspects of it.”