Chemically informed representations of amino acids enable learning beyond the canonical protein alphabet
J. V. CHRISTIANSEN, Miguel Gonzalez-Valdes Tejero, Chuang Sun Hembo, Yuchen Li, Carolina Barra
The paper introduces chemically informed peptide representations to enable machine learning models to learn from molecular structure rather than canonical amino acid symbols, improving prediction of MHC class I binding.
bioRxiv · Mar 16, 2026 · PreprintBioinformatics