MDDGlutar

MDDGlutar predicts protein lysine glutarylation sites using motif discovery and machine learning models derived from sequence features and statistical analysis of experimentally identified glutarylation sites.

Key Features:

• Glutarylation Site Prediction: Identifies lysine residues that may undergo glutarylation, a post-translational modification involving the addition of a glutaryl group.
• Motif Discovery with Maximal Dependence Decomposition: Uses maximal dependence decomposition (MDD) to partition glutarylation datasets into subgroups with conserved amino acid motif signatures.
• Sequence Feature Encoding: Extracts sequence-based descriptors including amino acid composition (AAC), amino acid pair composition (AAPC), and composition of k-spaced amino acid pairs (CKSAAP).
• TwoSampleLogo Motif Analysis: Applies TwoSampleLogo to identify amino acid enrichment patterns surrounding glutarylated lysine residues.
• Positional Dependency Analysis: Uses chi-squared tests to evaluate statistical dependencies between amino acid positions flanking glutarylation sites.
• Support Vector Machine Prediction Model: Integrates sequence features and motif information into a support vector machine (SVM) model for glutarylation site classification.

Scientific Applications:

• Post-Translational Modification Analysis: Predicts lysine glutarylation sites to support studies of protein post-translational regulation.
• Proteomics Research: Assists in identifying candidate glutarylated residues for experimental validation in proteomic datasets.
• Protein Function Investigation: Supports analysis of regulatory mechanisms involving lysine glutarylation in biological processes.

Methodology:

MDDGlutar analyzes experimentally identified glutarylation sites using TwoSampleLogo motif analysis and chi-squared tests, partitions sequence motifs with maximal dependence decomposition (MDD), encodes sequence features including amino acid composition (AAC), amino acid pair composition (AAPC), and composition of k-spaced amino acid pairs (CKSAAP), and applies a support vector machine (SVM) model evaluated using five-fold cross-validation and independent test datasets.