predML-Site

predML-Site predicts multiple lysine post-translational modification (PTM) sites within peptide sequences to identify acetylation, crotonylation, methylation, and succinylation for computational proteomics and related biological research.

Key Features:

• Multi-PTM prediction: Predicts multiple lysine PTM sites in peptide samples that may contain single, multiple, or no modifications.
• Supported PTM types: Targets lysine acetylation, crotonylation, methylation, and succinylation.
• Sequence encodings: Represents peptides using sequence-coupling, binary encoding, k-spaced amino acid pairs, and amino acid factorization.
• Feature selection: Integrates ANOVA (Analysis of Variance) and incremental feature selection to identify informative features.
• Classifier: Employs a cost-sensitive Support Vector Machine (SVM) classifier.
• Class imbalance handling: Uses cost-sensitive SVM to address class-label imbalance common in PTM datasets.
• Performance metrics: Reports 84.18% accuracy using the top 91 features, an aiming rate of 85.34%, and a coverage rate of 86.58%, reported to surpass existing predictors.

Scientific Applications:

• Computational proteomics: Aids identification and annotation of lysine PTM sites in proteomic datasets.
• Cell biology: Supports studies of PTM-driven regulation of cellular processes.
• Pathogenesis: Facilitates investigation of PTM roles in disease mechanisms.
• Drug development: Informs target characterization and modulation involving lysine PTMs.

Methodology:

Features are encoded with sequence-coupling, binary encoding, k-spaced amino acid pairs, and amino acid factorization; ANOVA and incremental feature selection identify top features; a cost-sensitive SVM classifier is used to predict K-PTMs while addressing class-label imbalance.