When: Feb 8th, 2023 – 11:00 – 11:30 AM
Where: Google meet link
Description
FEENI – FEature Encoding for Neoepitope-induced Immune response prediction
In collaboration with Valeria Repetto and Danilo Tatoni from CTGLab (CNR-IIT Pisa)
In this work, we are aiming to predict immunogenicity of a given Neoepitope, small peptide sequences capable of inducing a CD4+ response, using Deep Learning techniques.
This is a peculiar problem, since immunogenicity is not an intrinsic quality of the neoepitope, but is a result of its binding with the human leukocyte antigen (HLA). Furthermore, this response is divided in two steps: epitope binding with the Major Histocompatibility Complex (MHC) and the subsequent T-cell activation.
Since this is a work in progress, as of now we are attempting to encode the small (8-11 peptides) epitope sequences, described with 11 physicochemical properties, with a Fully Convolutional AutoEncoder (FCAE). The resulting encoding will then be used to train a classifier capable of predicting the immunogenicity of given epitopes.