Patchwork

Patchwork retrieves and concatenates phylogenetic markers from genome skimming and fragmented short-read assemblies to recover single-copy nuclear and organellar coding regions for phylogenomic analyses.

Key Features:

• Genome skimming support: Targets low-coverage whole-genome sequencing ("genome skimming") datasets for marker recovery.
• Input types: Mines phylogenetic markers from highly fragmented short-read assemblies and directly from sequence reads.
• Alignment engine: Uses DIAMOND for sequence similarity searches to identify homologous regions.
• Implementation: Implemented in the Julia programming language for computational performance.
• Hit stitching: Merges adjacent or overlapping homologous regions into contiguous units.
• Sliding-window trimming: Applies a sliding window algorithm to trim noncoding regions and retain coding portions.
• Marker recovery: Recovers near-universal single-copy orthologs, including single-copy nuclear markers and organellar coding regions.
• Performance: Demonstrated higher accuracy and computational speed than other programs in benchmarking across different sequencing depths.

Scientific Applications:

• Phylogenomic marker extraction: Recovery of single-copy nuclear and organellar coding sequences from genome skimming data for large-scale phylogenetic analyses.
• Marker mining from fragmented assemblies: Extraction of homologous phylogenetic markers from highly fragmented short-read assemblies and reads.
• Comparative benchmarking: Evaluation and comparison of marker-recovery performance across sequencing depths and against other tools.

Methodology:

Patchwork performs sequence similarity searches with DIAMOND to identify homologous regions, stitches adjacent/overlapping hits into contiguous units, and applies a sliding-window trimming algorithm to remove noncoding regions; it is implemented in Julia and accepts fragmented assemblies or raw reads as input.