Find biological associations between ‘never thought before to be linked’ entities
There was a time when science depended on manual efforts by scientists and researchers. Then, came an avalanche of data and with it the impossibility to read, annotate, and analyze hundreds of articles a day and find required connection between relevant entities. It is rather easy to say that a disease can be cured by a drug, or that a drug affects a protein, but finding associations between loosely connected entities or those which are two or three degrees of separation away is not always what one sets out to research. Indirect findings are like a cloud of silver linings that offer hope for acceleration of drug discovery and drug repurposing efforts.
In order to treat a disease, the researchers must understand all of the molecular pathways that lead to it. Understanding of these molecular pathways and disease mechanism is developed piecemeal by numerous researchers at different research institutes across geographies and time (decades or longer). The largest challenge for the researchers is joining together all of the information to derive a complete pathway that controls a disease. This is typically done by reading hundreds of scientific papers and analyzing all of the interactions that underpin a disease or biological process. Artificial intelligence solves this issue with natural language processing to point out specifically which articles can provide important and accurate insights about the selected entity. However, the relevancy of the context to the needs of the researcher remains unspecified.
Use of AI and network analysis in life sciences
With the use of AI in pharma, the industry has unlocked opportunities for analyzing data better and generating deeper insights from trillions of data points. However, even after vast repositories of life science data are analyzed, important links and biological connections between entities are missed or overlooked. This suggests that merely the availability of assets is not enough, but the right use of those resources is also necessary. Bringing together the world’s scientific and medical data is not enough, if done without catching on loose ends. Enabling network analysis of previously thought to be unconnected entities and the visualization of associations between diseases, genes, pathways, and drugs, especially those which are 2 degrees away, and could bring revolutionary insights to the pharmaceutical industry.
Discovery revolves around deep analysis and repurposing involves innovation. So, what could be an innovative way to use AI in order to enable and accelerate drug discovery and drug repurposing efforts? Innoplexus answers this question with one product: OntosightⓇ Explore.
OntosightⓇ Explore – Visualizing ‘never thought to be linked’ entities
OntosightⓇ Explore helps a researcher understand the interconnectedness of biological systems with regard to their search term, which can be a gene, target/protein, pathway, or disease. The module aggregates all of the diseases with their associated pathways and a series of molecular interactions which are responsible for its origin and severity. As, target identification is one of the most crucial steps in drug discovery which opens a door for new therapeutic development. Researchers can leverage the OntosightⓇ Explore module to identify all known and potential pathways through which a treatment breakthrough can be achieved where there are viable market opportunity.
The entire drug development journey is almost ~12-15 years which include lead identification to clinical development. Drug repurposing is an approach through which this time and clinical development cost can be minimized. Traditionally, researchers put huge efforts to access literature and identify alternative indications for the given drug. However, using OntosightⓇ Explore alternative indications for given drugs can be identified through indirectly associated indications through alternative targets and pathways. OntosightⓇ Explore also allows users to rank these associations and prioritize assets based on their commonality/association, as well as their druggability and druglikeness.