Photo by The Power of Large Language Models in Animal Study Results
As the field of artificial intelligence continues to evolve, we’re seeing a surge in AI-led biotechs that are revolutionizing early drug discovery. In this article, we’ll delve into the world of Large Language Models (LLMs) and explore how they’re being used to improve animal study results.
A example of AI-led biotech in early drug discovery
In our previous article, we discussed the role and challenges of LLMs in early drug discovery. Today, we’ll provide a practical example of how ChatGPT can be used to enhance scientists’ daily tasks.
A Simple Case Example
Our example demonstrates the power of LLMs in improving study results. We’ll use extracted measurements from 10 papers on acarbose-treated mice to provide recommendations on dose, participants, and measurements based on the results of the primary study.
Goal: Using extracted measurements from 10 papers on acarbose-treated mice to improve the recommendations made from the results of the primary study.
The key outputs required from this example include recommendations on dose, participants, and measurements, along with supporting data points. However, one of the challenges faced in implementing this example is creating prompts to accurately extract information to support recommendations.
The Importance of Prompt Engineering
To get the most out of ChatGPT, we need to employ some ‘prompt engineering.’ This involves writing prompts in a way that accurately describes the content of multiple files and papers. For instance, the first prompt in this example provides background and context to ChatGPT:
Prompt: “You are a drug discovery scientist looking to make decisions on dose, participants, and measurements when taking an existing diabetes drug into the ageing-related diseases field. You have experimental results from a mouse study that show the effects of acarbose on lifespan, body weight, body composition, fat pads, glucose, grip strength, grip duration, rotarod, and pathology. You also have several relevant scientific publications with studies investigating the effects of acarbose on different measurements in mice. You now want to interrogate your study results (which are in Excel files and images) and the publications separately for insights, and then together to get the best set of recommendations for your colleagues who are looking to perform early clinical trials with acarbose on ageing-related diseases.”
Results from the last prompt
As you can see from the screenshot above, ChatGPT has provided some valuable insights. However, there are some nuances that it hasn’t picked up on. For example, in female mice, the lifespan is not extended as much compared to male mice, but their physical measurements are improved. Improved prompts will aid the generation of more nuanced results.
Conclusion
In this article, we’ve demonstrated the power of LLMs in improving animal study results. By using ChatGPT to extract measurements from multiple papers, we can provide more accurate recommendations on dose, participants, and measurements. However, it’s essential to remember that LLMs have their limitations, and prompt engineering plays a crucial role in getting the most out of these models. We’d love to hear about your experiences with using LLMs in early drug discovery. Share your findings with us in the comments below!
Next in the Series
In our next article, we’ll discuss the key challenges in the effective use of LLMs for early drug discovery and present some practical approaches to address them.
