ImRex - CNN

ImRex (Interaction Map Recognition) is an convolutional neural network approach to predict the recognition between a T-cell receptor (TCR) and an epitope. It relies on a novel feature representation method that converts the amino acid sequences of the TCR and epitope into an interaction map, which combines the physicochemical properties of both interactors.

For the associated publication, see:

Moris, Pieter, Joey De Pauw, Anna Postovskaya, Sofie Gielis, Nicolas De Neuter, Wout Bittremieux, Benson Ogunjimi, Kris Laukens, and Pieter Meysman. “Current Challenges for Unseen-Epitope TCR Interaction Prediction and a New Perspective Derived from Image Classification.” Briefings in Bioinformatics 22, no. 4 (2020). [https://doi.org/10.1093/bib/bbaa318(https://doi.org/10.1093/bib/bbaa318).