Precision Medicine & Multiomics Data in Target & Drug Discovery
Activating Insights through Innovation
Humans beings are different from each other, and these differences run deep. They are present at a molecular level – on DNA, RNA transcripts or proteins. Knowing the molecular differences in a single person when they are healthy, compared to when they are ill, gives us a unique profile of molecules that can be used to diagnose and treat disease. Understanding the molecular profiles and differences between people, or groups of people, can help us design more precise and personalized medicines for a broader population.
Historically, these differences or variations were studied in the context of a specific gene, RNA transcript or protein between different people. With the advent of new technological breakthroughs, we can study multiple genes, RNAs and proteins in different people. We’ve moved from an analysis of a single gene to many genes or genomes, single RNA or transcript to many RNAs or transcriptomes, and single protein to many proteins or proteomes. This is the field of multiomics.
So how can precision medicine and multiomics data be used across the drug-development process and how do today’s technologies play a role in acceleration?
Target Discovery
Once an unmet need has been identified and the market verified, the search for a druggable target (DNA, RNA or protein) begins. Basic research to identify druggable targets often takes place at universities and hospitals, can last several years and results in tens of thousands of published articles every month. Large language models (LLMs) can help to identify potential targets by scanning the literature for relationships in the multiomics data on sites such as PubMed and ClinicalTrials.gov.
This store of multiomics data can be combined with internal proprietary multiomics data and then mined to find potential drug targets and pathways using Generative AI. Non-specific generative pretrained transformers (GPT) such as Chat GPT4 and EPAM’s AI DIAL (as well as the more biomedical research-specific BioGPT, BioBERT and ScieBERT) may be used for mining large volumes of research data as well as writing and reviewing scientific articles.
Target Identification
Starting from genome-wide association studies (GWAS), single-nucleotide variations can be linked to a specific disease or trait – information that can then be used to implicate genes in specific cellular pathways that are important to the disease.
Knowing the proteomics data through protein-wide association studies (PWAS) and linking it to GWAS data, it is possible to connect specific variations to effects on protein expression in a specific disease.
Target Validation
Showing how a drug binds to its target and the cellular pathways or networks it triggers is an important and complex step in drug development, involving the use of several complex computational modelling and biophysical methods (such as nuclear magnetic resonance, protein crystallography, binding assays or cryo-electron microscopy). Integrating the multiomics data from discovery and identification gives a system-level view of drug binding and the triggering of structural changes that lead to activation of specific intracellular pathways.
Having the appropriate data handling and viewing capabilities not only helps to streamline this process, but also allows for proper storage of the data using the principles of FAIR (Findable, Accessible, Interoperable and Reusable). This data needs to be stored with the right metadata and maintain the correct ontologies so it can be easily located, reused and be interoperable with data from other research groups.
Once that multiomics data is integrated and used widely, it’s possible to get a holistic view of the relationship between genes, RNA transcripts, proteins and metabolomic pathways within cells. In this way, one can validate drug targets that may be missed in a mono-omic study.
Hit Generation to Lead Identification & Optimization
To validate a target, it is ideal to view the 3D structure of the target and understand how a drug molecule could bind to this structure in silico – a process commonly known as structure-based drug design. We use protein folding and viewing software to perform molecular docking and modelling, with the ability to view how molecules fit into the binding pockets of a 3D protein. Such an approach can screen large libraries of possible drug candidates virtually, providing the ones most likely to bind to the drug target and elicit modulation. Employing a combination of high-throughput experimental screens and in silico modelling, it is possible to prioritize the best lead candidates for drug development.
Today’s Tech, Tomorrow’s Medicine
We’ve long recognized the vast and multidimensional differences between human beings, but today’s tech landscape is enabling us to apply this knowledge it in groundbreaking ways. Weaving multiomics data throughout the steps of drug-development, researchers are uncovering intricate biological insights that drive more precise drug discovery and personalized medicine.
AI is playing a pivotal role in this transformation, analyzing massive datasets with unparalleled speed and accuracy, extracting patterns and predicting therapeutic outcomes that might otherwise remain hidden. AI models are accelerating R&D processes by identifying promising drug candidates, optimizing molecule designs and forecasting potential side effects before clinical trials even begin.
Specialized assets and accelerators further enhance these capabilities. Innovations in cloud are providing life science and healthcare organizations with user-friendly, secure and GxP-ready environments for analyzing and correlating omics data. Opensource solutions and cheminformatics tools are democratizing drug design, fostering collaboration in the scientific community. Together, these advancements are paving the way for a new era of precision medicine, where treatments are tailored to meet the unique biological and molecular needs of individual patients.
With deep domain knowledge, bespoke solutions and a collaborative approach, EPAM stands ready to partner with pharma and biopharma clients to advance their drug discovery and development efforts. Learn more about life sciences at EPAM.
