The application of deep learning to predicting events in the antigen processing and presentation pathway
Alex Rubinsteyn, Tim O'Donnell, Bulent Arman Aksoy, Nandita Damaraju, Giancarlo Kerg,
Icahn School of Medicine at Mount Sinai
Our lab has been working for several years to bring modern neural network software and methods into immunology. Our peptide/MHC class I binding affinity predictor, MHCflurry, has performed well on the IEDB’s automated benchmark since mid-2015. We’ve recently extended our work to predicting class II binding affinity as well as class I cleavage, stability, and presentation. Our code and models are open source without restrictions on commercial use. In this talk, we’ll share lessons learned from the MHCflurry project, discuss how this software can be used to answer basic and translational research questions, and outline our plans for future work in this domain.
Credits: None available.
You must be logged in and own this product in order to post comments.