Large Language Models in Drug Discovery and Development: From Disease Mechanisms to Clinical Trials

TL;DR

This paper aims to provide a comprehensive overview for researchers and practitioners in computational biology, pharmacology, and AI4Science by offering insights into the potential transformative impact of LLMs on drug discovery and development.

Abstract

The integration of Large Language Models (LLMs) into the drug discovery and development field marks a significant paradigm shift, offering novel methodologies for understanding disease mechanisms, facilitating drug discovery, and optimizing clinical trial processes. This review highlights the expanding role of LLMs in revolutionizing various stages of the drug development pipeline. We investigate how these advanced computational models can uncover target-disease linkage, interpret complex biomedical data, enhance drug molecule design, predict drug efficacy and safety profiles, and facilitate clinical trial processes. Our paper aims to provide a comprehensive overview for researchers and practitioners in computational biology, pharmacology, and AI4Science by offering insights into the potential transformative impact of LLMs on drug discovery and development.

Figures (6)

• Figure 1: Large Language Models Shaping the Future Landscape of Drug Discovery and Development. In the past, each stage of drug discovery involved numerous manual tasks, which requires significant human effort and substantial resources. Nowadays, advancements in biotechnology, alongside the integration of AI and computer-aided in silico computation tools, have reduced the need of human labor and resources. However, we have yet to have a highly automated drug discovery pipeline, especially in the clinical trial phase, where trail design and matching are still mainly done by clinical practitioners. In the future, it is anticipated that the continued development of LLMs and their application in drug discovery will enable a highly automated drug discovery process.
• Figure 2: The two main paradigms of language models. Specialized language models are trained on specific scientific languages and are typically tailored for specific or a few science-related tasks. These models are used as tools to perform a specific task, in which users provide the information required for a task, and the model outputs the prediction. General-purpose language models are trained on diverse textual information sourced from various materials, including scientific papers and textbooks. These models are used like an assistant that allows users to use plain language to interact with the model.
• Figure 3: Understanding Disease Mechanisms. The left part of the figure illustrates the processes involved in understanding disease mechanisms. This process involves clinical sub-typing, target-disease linkage analysis, and target validation. Clinical subtyping refers to identifying subgroups of patients with similar clinical characteristics during which data can be collected from multi-omics. Target-disease linkage analysis refers to identifying the relationship between targets and diseases. Target validation typically involves three steps: safety and feasibility, mechanisms of action, and modality selection. The right part of the figure highlights the tasks that LLMs can perform to facilitate these processes, including genomics analysis, RNA analysis, pathway analysis, target profiling, strategic profiling, and assistance.
• Figure 4: Drug Discovery. The left part of the figure illustrates the processes involved in drug discovery. The right part of the figure highlights the tasks that LLMs can perform to facilitate these processes.
• Figure 5: Clinical Trials. The left part of the figure illustrates the processes involved in clinical trials. Clinical trials consist of four phases: Phase 1, Phase 2, Phase 3, and Phase 4. The right part of the figure highlights the tasks that LLMs can perform to facilitate these processes.