Paolo Marcatili

The paper analyzes variance sources in LC-MS data and evaluates transfer learning's impact on ML performance across public datasets.

BepiPred-3.0 improves B-cell epitope prediction using protein language models, achieving high accuracy for both linear and conformational epitopes.

Combining an ESM-1b encoder with a conditional random field decoder, DeepTMHMM classifies alpha-helical and beta-barrel transmembrane proteins and their topology, outperforming most compared tools, including on signal-peptide cleavage sites.

The paper analyzes antibody sequences from patents to assess their utility for therapeutic antibody engineering and finds they reflect clinical sequences.

proABC-2 predicts antibody-antigen contacts using a convolutional neural network and integrates with HADDOCK for docking.

NetSurfP-2.0 improves prediction of protein structural features with higher accuracy and faster runtime than previous versions.

A sequence-only convolutional network trained on T cell receptor (TCR)-peptide pairs for HLA-A*02:01 can single out TCRs that bind a target peptide among many non-binding ones.

NetMHCpan-4.0 improves peptide-MHC class I interaction predictions by integrating binding affinity and eluted ligand data, enhancing identification of naturally processed ligands and T cell epitopes.