MAFFT-DASH

MAFFT-DASH integrates structural alignments from the DASH database with the MAFFT multiple sequence alignment program to improve MSA accuracy, particularly for remote homologs and cases of weak sequence similarity.

Key Features:

• Structural integration: Incorporates structural alignments into the MAFFT alignment process to supplement sequence information.
• DASH database: Utilizes a Database of Aligned Structural Homologs (DASH) providing structural alignments at domain and chain levels for proteins in the Protein Data Bank (PDB).
• Programmatic access: Exposes DASH structural alignments via a REST-like API for integration with computational workflows.
• Improved accuracy: Delivers a reported 10–20% improvement in alignment accuracy compared to standard MAFFT on benchmark datasets.
• Benchmarking and metrics: Performance evaluated on BAliBase, HomFam, OXFAM, Mattbench, and SISYPHUS using the Sum-of-Pairs (SP) metric.
• Remote homology detection: Leverages structural context to detect remote homologs that may be missed by sequence-only methods.
• Template library assembly: Supports assembling template libraries for homology modeling using structural alignments.

Scientific Applications:

• Multiple sequence alignment refinement: Improves MSA quality for datasets with low sequence similarity.
• Functional annotation: Aids detection of remote homologs to inform functional annotation of proteins.
• Homology modeling: Provides structural templates and assembled libraries for homology-based structure prediction.
• Structural biology and protein engineering: Supports structural inference and template selection for protein engineering studies.

Methodology:

MAFFT-DASH integrates DASH structural alignments (domain- and chain-level) with MAFFT; DASH is queried via a REST-like API; performance was benchmarked on BAliBase, HomFam, OXFAM, Mattbench, and SISYPHUS using the Sum-of-Pairs (SP) metric.