Metacoder

Machine Learning in commercial drug development is dependent on dataset size, quality, and software programming. Dataset preparation and integration are challenging when dealing with TBs of heterogeneous data across all multiomics. Skilled personnel in software, data engineering, and biochemistry are all in shortage and this is slowing the progress of treatment discoveries. To alleviate these challenges Metacoder offers UI-based machine learning which can deliver SOTA predictive and generative models for pharmaceutical research. It does this by interfacing with the world of public multiomic data sources like Chembl, BindingDB, TTD, OpenTarget, PharmGKB, Reactome, IUPHAR, and Stitch, and combining/preprocessing these sources into a useable root state. This root data can be combined with private data. Metacoder runs on a 1000 GPU computing cluster, so even the deepest models will run in reasonable time. Untold therapeutic treatments are dependent on machine learning and our goal is to improve access to that.