The goal of this project is to compare different strategies for predicting MHC-I binding peptides.For this reason, we have built several models that we put to test through a detailed comparative analysis. The models developed ascribe to the course curriculum on the subjects of position specific scoring matrix (PSSM), stabilisation matrix method (SMM) and artificial neural networks (ANNs). Several universal techniques have been incorporated for model training and validation, such as, optimizing algorithms, clustering of the peptides accounting for redundancy and a nested cross validation.

s161053 Konstantinos Kalogeropoulos

s143830 Hannah-Marie Martiny

s133583 Marius Thrane Ødum

s193364 Frederik Teudt

s200177 Athanasios Pasias