SeqDivA

SeqDivA detects remote homologs using graph theory-derived sequence and structure descriptors combined with alignment-free methods to improve detection in the twilight zone of low sequence similarity.

Key Features:

• Graph Theory-Based Descriptors: SeqDivA uses graph theory-derived sequence and structure descriptors to encode biological sequences for detection of remote homologs in diverse gene and protein families and superfamilies.
• Alignment-Free Methodology: SeqDivA implements alignment-free (AF) approaches as alternatives to alignment-based (AB) algorithms for detecting remote homologs in regions of low sequence similarity (twilight zone).
• Integration with Alignment-Based Features: SeqDivA supports integration of AF features with AB features within a single predictive model or by combining predictions using voting and weighting strategies.
• Scalability for Genomic Comparisons: SeqDivA scales AF and AB features and measures to enable comprehensive comparisons across multiple genomes and proteomes.

Scientific Applications:

• Comparative Genomics and Proteomics: Detecting remote homologs to infer evolutionary relationships across genomes and proteomes.
• Functional Annotation of Distant Homologs: Identifying subtle sequence similarities indicative of distant evolutionary connections to support functional annotation.

Methodology:

SeqDivA employs alignment-free approaches and graph theory-derived descriptors and builds on methodologies such as MARCH-INSIDE, TI2BioP, and ProtDCal, with integration options for traditional alignment-based methods.