BLAST

BLAST identifies regions of local similarity between nucleotide or protein sequences and sequence databases and computes statistical significance of alignments to infer homologous relationships.

Key Features:

• Local alignment (MSP): Approximates local alignments using the maximal segment pair (MSP) score to optimize local similarity.
• Sequence types and databases: Compares nucleotide and protein query sequences against sequence databases.
• Statistical scoring: Calculates statistical significance for alignments to assess the relevance of similarity hits.
• Heuristic speed: Uses a heuristic approximation that yields roughly an order-of-magnitude speed improvement while maintaining comparable sensitivity.
• Multiple-region detection: Detects and reports multiple similarity regions within long DNA sequences.
• Comparative sensitivity: Exhibits lower sensitivity for detecting remote homologs compared with profile-based methods such as CS-BLAST and PHMMER.

Scientific Applications:

• DNA and protein database searches: Identify homologous sequences by searching nucleotide or protein queries against sequence databases.
• Motif identification: Detect conserved motifs within queries via local alignments.
• Gene discovery: Support gene identification and annotation through sequence similarity.
• Analysis of long sequences: Characterize multiple similarity regions across long DNA sequences.
• Functional inference: Infer functional relationships between sequences based on alignment similarity and statistical significance.
• Remote homolog detection (comparative): Detect remote homologs with less sensitivity than profile-based approaches such as CS-BLAST and PHMMER.

Methodology:

Approximates alignments by optimizing local similarity through the maximal segment pair (MSP) score, computes statistical significance of alignments, and employs a heuristic approach that delivers substantial speed improvements while retaining comparable sensitivity.