MethK

MethK predicts methylated lysine residues on histone and non-histone proteins and characterizes protein methylation, a post-translational modification predominantly occurring on lysine and arginine and implicated in transcriptional regulation.

Key Features:

• Predictive Models: Two support vector machine (SVM) models are trained separately on histone and non-histone lysine-methylated data using features including amino acid composition (AAC) and accessible surface area (ASA).
• Model Specialization: Separate models for histones and non-histone proteins account for differences in sequence conservation between these protein classes.
• Performance Metrics: Model performance is evaluated by five-fold cross-validation with reported histone sensitivity 85.62% and specificity 80.32%, and non-histone sensitivity 69.1% and specificity 88.72%.
• Flanking Region Analysis: Characterizes features of the flanking regions surrounding lysine-methylated sites in both histone and non-histone proteins.
• Functional Analysis: Conducts gene ontology (GO) functional analysis of lysine-methylated proteins.
• Correlation Studies: Analyzes correlations between lysine-methylated sites and other post-translational modifications (PTMs) in histones.

Scientific Applications:

• Site Prediction: Predicting lysine methylation sites on histone and non-histone proteins.
• Comparative Analysis: Distinguishing methylation patterns between histones and non-histone proteins based on sequence conservation.
• Motif and Context Characterization: Characterizing sequence and structural features of flanking regions around methylated lysines.
• Functional Annotation: Assigning gene ontology-based functions to lysine-methylated proteins.
• PTM Crosstalk Investigation: Investigating correlations and crosstalk between lysine methylation and other histone PTMs.

Methodology:

Two support vector machine (SVM) models trained separately on histone and non-histone lysine-methylated data; feature set includes amino acid composition (AAC) and accessible surface area (ASA); model evaluation by five-fold cross-validation; additional analyses comprise flanking region characterization, gene ontology functional analysis, and correlation analysis between lysine-methylated sites and other PTMs.