GenAI and LLMs to potentially 'revolutionise' drug discovery - what top pharma thought leaders have to say

Mira Nair

On July 11, 2023, Dr Daniel Jamieson, CEO, Biorelate, chaired a virtual panel with thought leaders Dr Ashar Ahmad, Data Science & Advanced Analytics, Associate Director at Grünenthal, Dr Shameer Khader, Global Head of Data Science, Data Engineering and Computational Biology at Sanofi, Jon Hill, Principal Scientist, Boehringer Ingelheim, and Dr Andrew Davis, Director of Knowledge Management, at Novartis.

To get beyond the universal hype and buzz around ChatGPT, Generative Artificial Intelligence (AI), and Large Language Models (LLMs), ubiquitous across all industries, we brought together biopharma thought leaders to dive into what is hallucination and what is real. 

What could the applications of LLMs be for data science and bioinformatics? What change could they really bring? What are they key elements that need to be in place to ensure the transformative potential of LLMs is not a hallucination but something achievable over the next few years?

Watch the whole webinar recording here after you read some of the insightful take-aways below (ironically, we did not use ChatGPT in any part of the curation or writing of this blog post).

1. The applications of LLMs are practically limitless, but a few stand out as low-hanging fruit.

While target ID and adverse event data management and other immediate applications of LLMs, our experts opened our minds to the overarching value these technologies have for the field.

 “Generative AI and other tools are an evolution of the way we've dealt with text over the years. Recent developments in generative AI are impactful in providing the next step to something that's both more flexible and potentially more powerful.”- Jon Hill

“These new approaches are able to bring in a lot more value. We are able to see new biology which we are not able to see otherwise; we are in an era of generative biology.” - Dr Shameer Khader

“As a scientist or as any human being, when we think about stuff, when we reason about it, we are reasoning to some degree with the help of natural language. So, a lot of these ideas that these LLMs are a paradigm shift in fact come from this idea that language is somehow special, human language is somehow special, so if you can build a foundation model on natural language, emergent intelligence can come up just because language is so complex. If you can build certain foundational models around that, it can really be revolutionary.” - Dr Ashar Ahmad

“The moment is already here. Everybody now knows that they want to chat with your data set, nobody wants to search your data set anymore, so we see that the public perception on this has changed, the way we interact. And scientists want us, like, yesterday, to have this type of capability ready to interact with their data.“ - Dr Shameer Khader

“One of the areas that I'm most excited about is really using these large language models to flesh out the landscape around new drug targets.“ - Jon Hill

“For myself and other bioinformaticians, we work so much directly with code in terms of data analysis, generating visualisations, things like this, so we can't underestimate how important the impact of large language models directly on code generation is going to be.“ - Jon Hill

2. The status quo of biomedical literature search is many years behind what it should be.

The potential of LLMs highlights the limitations of what scientists use now for querying the biomedical literature. The industry thought leaders highlighted the ‘tool fatigue’ most scientists deal with every day. 

“One of the most used tools by bench scientists today is still PubMed, but PubMed hasn't changed much over the last 30 years; even with all the new technology, it's still Boolean search.” - Dr Andrew Davis

“The next step is bringing in the context of the literature...Just saying, 'Here's all the papers on PubMed on this target, that should be all the information you need,' that's not really going to get them what they want, and it's really going to be the summarisation around the key aspects that's most important. ” - Jon Hill

“I'm hoping the first generation of the generative AI tools integrate well with existing tools and they become a true enhancement on those old tools rather than yet another tool to go to” - Dr Andrew Davis

“We need better foundation models in biomedicine. Industry and then technology and then pharma partners need to come together to develop foundation models to take what we can do to the next level.” - Dr Shameer Khader

“LLMs might help you also refine your hypothesis and design your hypothesis in an interactive fashion. So, not just a design idea, that then goes into an experiment, then the data gets generated, the machine learning just does the data analysis, but rather it starts at the very beginning, so how do these scientists iteratively think about going about their target hypothesis generation?” - Dr Ashar Ahmad 

3. Will LLM be the answer to most of the current problems for data science?

The key question biopharma wants answered is ‘Will LLM technology bring together all of the developments we’ve seen in biopharma data over the last 30+ years into one unified solution?’ Specific aspects of generative AI could kill many birds with one stone and bring together an overarching capability to help data scientists advance biological knowledge to speed drug discovery and drug design faster than any other time in the history of medicine.

“Where tools like LLMs will really enable change is in bringing everything together and then the scientists will be able to work in the same way that they think, they won't have to adopt a new language to query or ask a question.”- Dr Andrew Davis

“The moment is already here. Everybody now knows that they want to chat with your data set, nobody wants to search your data set anymore, so we see that the public perception on this has changed, the way we interact. And scientists want us, like, yesterday, this type of capability to interact with their data.” - Dr Shameer Khader

“Medicinal chemists in most companies now are using generative AI as co-pilots. And this is very important because machine learning and AI is seen by a lot of people as just a data analysis tool; there is limited understanding that these technologies can not only help us analyse the data, but also help us plan the experiment that can generate the data.” - Dr Ashar Ahmad 

By popular demand, we are hosting a sequel to this webinar on October 17th, 2023, for the biopharma community with the same thought leaders speaking from Novartis, Boehringer Ingelheim, and Grünenthal. 

Register your interest in attending here; a recording will be sent out to all approved registrants after the event, even if you aren’t able to attend on the day: GoTo Webinar

Or follow us on LinkedIn for the latest updates on what is possible for biomedical literature search with the latest LLM, AI and NLP technology: