Predicting Antibody Properties from Sequence using Machine Learning

Is it possible to predict properties of antibodies from sequence using machine learning? That's what a group of students at Harvey Mudd College are trying to do! Using Pipeline Pilot, the members of BIOVIA's Clinic team are building models to predict an antibody's Developability Index (DI; Developability Index: A Rapid In Silico Tool for the Screening of Antibody Aggregation Propensity) from sequence rather than 3D structure. This would allow scientists to more quickly screen candidate antibodies as the costly and time consuming step of predicting or determining the 3D structure of the antibody can be skipped. Attached is there mid-year presentation that they gave discussing their approach to solving this problem.


​​​​​​​BIOVIA's 2019-2020 Harvey Mudd Clinic Team Mid-Year Presentation