AI- Powered Target Identification for Disease Indications : Integrated Approach Encompassing Diverse Types of Dataset
Objective
- Target identification is the process of identifying the correct biological molecules (protein, gene or RNA) or cellular pathways that can be modulated by drugs to achieve therapeutic benefits. Target discovery is the most crucial initial step in drug discovery that influences probability of the overall success of drug development. The lack of clinical efficacy and safety are the major contributing factors for failure of drugs in clinical development emphasizing the high importance of selection of right target for drug discovery programs.
- Traditional target discovery encompasses studying of plethora of research publications to identify the genes and proteins involved in novel disease mechanisms. These efforts are usually followed by identifying the “druggability“ of the target by assessment of multiple features of the target protein. With application of various AI and ML methods, using all publicly available scientific publications & Omics –datasets, we aim to reduce the time taken for this time-consuming process and increase the likelihood of the identification of successful targets for disease indication.
Methodology
We apply a combination of diverse types of artificial intelligence-based algorithms on various datasets including publicly available research articles and various disease specific Omics- datasets.
- Generate knowledge graphs and utilize machine learning models and integrate various gene expression datasets for data integration.
- Identify promising proteins with the required evidence level in disease pathogenesis.
- Assess proteins' amenability to different drug modalities (e.g., small molecules, biologics).
- Ensure effective target identification and eliminate undruggable targets
- Use tissue-specific target expression profiles to exclude potentially unsafe targets.
- Understand target function and signaling to evaluate potential on-target adverse effects.
- Identify novel disease-correlated genes and genes with robust molecular-level evidence.
- Assess the target for the disease in the context of entire landscape of approved and “in-development” pipeline.
- Leverage competitive intelligence data for comprehensive target analysis.
Outcomes / Impact
Our AI powered tailored target identification approach enables the rapid identification of top disease associated target genes reducing overall time taken and expenditure. This methodology also helps to identify the novel as well as repositionable targets for any therapeutic indication. The integrated customized process enables the user to select different key parameters such as druggable features, novelty level, potential on-target liabilities etc. required for target discovery. Our AI algorithms help to efficiently integrate vast bodies of scientific literature, databases and diverse omics- datasets specific for disease ensuring all the disease relevant mechanisms are considered for target discovery.