Disease modelling and target identification are critical and time-consuming portions of the drug discovery process and areas where artificial intelligence (AI) is already making a substantial impact.
A new opinion paper in the Cell Press journal Trends in Pharmacological Sciences traces the history of target discovery methods – from experimental approaches, to multi-omics approaches, to machine learning and AI methods – and details how companies are using cutting-edge AI technologies to find targets and develop drugs now in phase I and phase II trials for diseases including cancer, Alzheimer’s disease, diabetes, COVID-19, and idiopathic pulmonary fibrosis.
“AI has ushered in a new era of target identification and drug discovery, and has allowed us to more quickly and efficiently find new pathways to treat difficult diseases,” says Frank Pun, PhD, Head of Insilico Medicine’s Hong Kong office and lead author on the paper.
As the paper notes, scientists have long relied on experimental methods like stable isotope labelling by amino acids in cell culture (SILAC) to uncover new targets. SILAC, a type of comparative profiling, uses stable isotope-labelled amino acids to accurately differentiate cellular proteomes, uncovering key disease pathways for numerous cancers, including hepatocellular carcinoma, multiple myeloma and colorectal cancer.
CRISPR-Cas9 gene editing has also served as an important tool, expanding scientists’ understanding of disease mechanisms including in COVID-19 and multiple myeloma.
And scientists have uncovered hundreds of thousands of associations between genetic variants and complex diseases using the genome-wide association study (GWAS).
This method has led to breakthroughs in therapies for cystic fibrosis and inflammatory bowel disease among other conditions.
The amount of biomedical data available to researchers – from basic research to clinical investigations of patients – has grown substantially in recent years, posing added complexity for human scientists, but perfect conditions for AI which can quickly process and analyse complex networks of data.
“AI is a powerful tool in the target discovery and drug development process,” says Alex Zhavoronkov, PhD, founder and CEO of Insilico Medicine and co-author of the paper, “and its role in drug development continues to expand. We envision a future where AI plays an indispensable role in accelerating the development of safe and effective therapeutics.”
AI tools are already assisting scientists in biomarker and target identification, indication prioritisation, drug-like molecule design, pharmacokinetics prediction, drug–target interaction, and clinical trial design.
The paper highlights 10 biotechnology companies using AI tools to develop new therapeutics that have reached phase I or phase II trials.
Insilico Medicine’s founder and CEO Dr Zhavoronkov is a pioneer in developing generative AI algorithms for target discovery and novel drug design and the company’s end-to-end platform, Pharma.AI, has produced three novel drugs in clinical trials, including a lead drug for idiopathic pulmonary fibrosis that is in phase II trials, the first fully generative AI drug (with both an AI-discovered target and AI design) to reach this stage.
Insilico has 30+ drugs in its pipeline, and has nominated 12 preclinical candidates since 2021 for indications including cancer, fibrosis, immunity, central nervous system diseases, infectious diseases, autoimmune diseases, and ageing-related diseases.
The paper notes many ways AI has accelerated and improved the drug discovery process, including generating synthetic data where therapeutic data is scarce and to help address data imbalance or bias issues.
Importantly, the paper describes how AI can also help researchers zero in on the best target, including identifying causal inference between targets and disease, and using natural language processing to extract evidence from publications, grants and other data sources to further refine a target’s novelty and likelihood of success.
Source: Insilico Medicine