HiFiX

HiFiX performs high-fidelity clustering of homologous protein sequences to separate highly divergent homologs and resolve families affected by protein domain rearrangements.

Key Features:

• Comprehensive Clustering Strategy: Clusters sequences based on entire-sequence length and gene-family-specific substitution patterns to increase sensitivity and accuracy.
• Pairwise Similarity and Pre-family Formation: Performs pairwise sequence comparisons using permissive similarity criteria to group sequences into pre-families.
• Homogeneous Cluster Refinement: Subdivides pre-families by analyzing the topology of the similarity network to obtain more homogeneous clusters.
• Family Merging and Model Selection: Computes multiple sequence alignments and applies model selection techniques to merge clusters into statistically supported families.
• Robustness to Sequence Divergence: Adapts to sequence divergence and protein domain rearrangements to distinguish highly divergent homologs from partial homologs.
• Scalability: Designed to operate on very large datasets while maintaining clustering accuracy and computational feasibility.

Scientific Applications:

• Genomics: Delineates gene families to support analyses of gene content and family-level evolution across genomes.
• Proteomics: Facilitates grouping of protein sequences for studies of protein function and functional divergence.
• Phylogenetics: Supports inference of phylogenetic relationships and the functional evolution of genes through robust family definitions.
• Protein domain architecture analysis: Identifies and separates sequences affected by domain rearrangements and partial homology for domain-centric studies.

Methodology:

Performs pairwise sequence comparisons with permissive similarity criteria to form pre-families; analyzes similarity network topology to subdivide pre-families into homogeneous clusters; computes multiple sequence alignments and applies model selection to merge clusters into final families.