MetaCRAST

MetaCRAST detects CRISPR arrays and spacer sequences in raw, unassembled metagenomic reads by using expected direct repeat (DR) sequences derived from assembled contigs or bacterial genomes as references.

Key Features:

• Reference-Guided Detection: Uses expected CRISPR direct repeat (DR) sequences derived from assembled contigs or bacterial genomes to guide detection in raw, unassembled metagenomes.
• Wu-Manber Multipattern Search: Employs a fast implementation of the Wu-Manber multipattern search algorithm to identify metagenomic reads containing expected DR sequences.
• Spacer Extraction and Clustering: Locates DR sequences within acceptable distances to extract spacers and clusters extracted spacers into a non-redundant set using CD-HIT.
• Performance Evaluation: Evaluated against Crass and MinCED, showing superior spacer detection in real metagenomes from acid mine drainage (AMD) and enhanced biological phosphorus removal (EBPR), outperforming de novo methods on simulated Illumina metagenomes and yielding comparable results on simulated 454 metagenomes.
• Implementation and Parallelization: Implemented in Perl and supports parallelization via the Many Core Engine (MCE).
• Input Formats: Accepts metagenomic sequence reads and direct repeat queries in FASTA or FASTQ formats.

Scientific Applications:

• Viral–Host Linkage: Links environmental viruses to their microbial hosts by matching CRISPR spacer sequences to viral sequences.
• Microbial Ecology: Enables analysis of viral-host dynamics and microbial community interactions in diverse environments.
• Evolutionary Analysis: Supports investigation of the evolutionary history of microbial populations via CRISPR spacer records.

Methodology:

Uses expected DR sequences from assembled contigs or bacterial genomes, applies a Wu-Manber multipattern search to select DR-containing reads, extracts spacers by locating DRs within acceptable distances, clusters spacers with CD-HIT, and is implemented in Perl with optional Many Core Engine (MCE) parallelization; inputs are FASTA or FASTQ.