<ul><li>RankNovo is introduced as the first deep reranking framework for enhancing de novo peptide sequencing by combining multiple sequencing models' strengths.</li><li>RankNovo uses a list-wise reranking approach, models candidate peptides as multiple sequence alignments, and applies axial attention to extract informative features.</li><li>New metrics PMD (Peptide Mass Deviation) and RMD (Residual Mass Deviation) are introduced for quantifying mass differences between peptides at both sequence and residue levels.</li><li>Extensive experiments show that RankNovo surpasses base models, sets a new state-of-the-art benchmark, and generalizes well to unseen models, demonstrating its potential as a universal reranking framework for peptide sequencing.</li></ul>

Universal Biological Sequence Reranking for Improved De Novo Peptide Sequencing

Discover more