Going back to biology and the hype of AI in pharma
Posted 2nd August 2019 by Jane Williams
With a background in pharmacology, toxicology, molecular cell biology and genetics, Rangaprasad Sarangarajan joined BERG Health in 2010 as head of R&D.
In the initial years of the company, he aided in the conception, design, and implementation of the paradigm of looking at human biology, use of technology to generate molecular signatures, and the use of artificial intelligence-based analytics for understanding the complexity of biology, identifying of targets, and developing of drugs, and its utility in clinical development.
Berg has a focus on ‘Back to Biology’. Can you explain a little more about that and why it’s a priority for Berg?
Traditional biology has always been predicated on prior understanding of the nature of disease, identification of targets based upon those hypotheses, and trying to work towards a goal of developing therapeutics based upon pre-existing knowledge.
Those have been reasonably successful, but over the past few years, there hasn’t been any seminal findings in terms of additional understanding of pathways underlying diseases including cancer. For unravelling the complexities of biology, we realised there was a need to dig into the biological architecture.
The only way to do that was to holistically develop disease models based upon everything developed from human bio specimens. Once you start developing these models, there is so much data generated, so the only way to be able to analyse it reasonably is with the use of AI analytics.
Our back to biology paradigm is primarily walking away from any hypothesis-driven approach that exists within the scientific community and moving towards generating these disease models and enabling the data to generate reasonable hypothesis that could potentially be used for treating disease in a better manner than what currently exists.
Why is AI in pharma getting so much attention now?
From my perspective, there is a lot of hype portrayed in the public domain including the pharmaceutical industry, but there is a good reason to move into the AI space, especially given that the technology has advanced to a point where it is possible to generate really comprehensive datasets from biological systems that was not possible even five years ago. The last 10 years has seen such incredible advances in technology to generate data focused on teasing out the complexities of biology, and the only way to analyse that would be domain-specific analytics.
When we talk about AI in pharma, everybody thinks that one size fits all. What people don’t realize is that the AI has to be domain specific. For example, BERG started looking at building our own AI suite long before AI was the buzzword. We started looking at it early 2009 when AI was not a big buzzword in the pharma industry and have focused specifically on building an AI platform that is able to dig deep into the biology and provide meaningful, simple outputs that can then be leveraged for not only understanding the disease but identifying targets, moving this along for drug discovery and development. BERG now is actively using these components within the clinical development.
Because BERG was one of the early adopters of AI, were people initially cynical or hard to convince about its use?
Absolutely! There was more cynicism than anything else. I still remember the look on people’s faces as we were sitting in meetings saying we were going to look at biology and identify targets using AI and using data in AI. There was that sense of ridiculousness or incredulity. But as AI has become more prominent in the mainstream, those challenges have decreased.
Also, the significant volume of data we have as through points, produced over the past decade, is validation of our ability to leverage AI in the biological realm, especially with focus on pharma to try and not just critical unmet needs in terms of data-driven prioritization of targets for specific diseases.
Our focus has been on neurology, especially Parkinson’s disease and Alzheimer’s disease where we have had some good success in identifying novel targets that are now in drug development. We’ve also had success in areas with critical unmet need, in immunological diseases like lupus, where we already have some molecular signatures.
Was it ever difficult to keep going in the face of that cynicism?
We were all very invested in the vision of the company. It helped when we put the pieces together and got serial success when we repeated the process, especially in the scientific arena.
Yes, there were challenges as we moved along, but those helped us streamline the process to a point where we are so good in terms of our availability to build the biological models, defragment the biological models into their molecular components, use the AI to ask questions that are meaningful in terms of creating our pipeline and developing solutions for the critically-unmet need populations in the health sector.
What are your predictions for the use of AI in the future?
There are so many different flavours of AI that people are using in so many different things. From my personal perspective, the future is wide open, based upon the diversity of data that can be collected for every individual, not only from our perspective of building biological models and looking at molecular datasets, but also when you look at the peripherals, the sensors, and everything else being developed will add to the granularity of the signature of a single human being.
Target data collection is going to be so large that it will require AI of different flavours to be collectively used to make any meaningful sense. Like everything else, there is going to be a spurt of activities, those things that work will shine and those that don’t work will drop off like any other development cycle.
We are almost in the biological sectors, at least in my humble opinion, reaching a point, where those initiatives within the pharma industry where AI has the through point, companies like BERG will persist in our quest to reach the next scientific frontier in the integration of biology and AI. Others may not actually reach that point.
What are you most excited about for the near future for BERG?
I’m in a very enviable position I think within the pharma industry where I have foresight on one of the most diverse pipelines. Our team is working on targets that are first in class, which to the best of our knowledge, nobody else is working on.
The most wonderful part about this is these are targets that were derived from platform using a data-driven approach, but the novelty is not towards the commercial success of it. Instead its towards our ability to demonstrate the scientific basis that can truly make a difference in the patient’s life from diagnosis to trying to slow the progression of the disease, trying to prevent the disease, and hopefully enable cures in very specific areas, specifically some rare oncology indications as well as Parkinson’s disease, Alzheimer’s disease, and autism.
Rangaprasad Sarangarajan, Senior Vice President and CSO, BERG will be speaking on ‘AI for threading causative biology of disease to genetic stratification for clinical development – impact on time (& cost)’ at the upcoming 3rd Global Pharma R&D Informatics & AI Congress.
Interested in finding out more? Download the agenda for the 3rd Global Pharma R&D Informatics & AI Congress.
Leave a Reply