Dr. Alix Lacoste, VP Data Science at BenevolentAI and Dr. Michelle Gill, Senior Data Scientist at BenevolentAI, will present: "Machine Learning for Target Identification and Lead Optimization in Drug Discovery".
The talk will be held at JLABS@NYC, starting at 6pm. The networking will be a separate event held and hosted by NYAGIM at a nearby location, starting at 7pm.
A key challenge in the drug discovery process is the identification of potential therapeutic targets for a given disease. This process requires selecting both a target and an entity, often a compound, for modulating the target’s activity to validate its association with a disease. An additional complication is that heretofore unknown targets may be more challenging to identify but offer increased opportunities for the development of novel drugs.
Once a target is validated, a compound must be identified which modulates the target in the desired fashion and has desirable properties, such as solubility and low toxicity. At BenevolentAI, we use machine learning throughout the entire process. Relation extraction models drive our unstructured pipeline. In conjunction with structured data and our own experimental results, our processing pipeline ingests information from scientific publications, abstracts, and patents. Next, this biological knowledge graph is leveraged to form predictions
using relational inference algorithms, including matrix factorization and graph convolutional models. These predictions are aggregated together with models built on genomics data and information mined from text. Druggability, tissue specificity, and other metadata are then
added to surface the most promising targets to test in the lab. Models trained on chemistry data sources are able to predict relevant physicochemical properties, as well as automatically design molecules and predict synthesis pathways.
This talk will describe our process and highlight some of our early successes.