GraftM

GraftM classifies metagenomic sequences into gene families and phylogenetically places them into pre-constructed gene trees to provide taxonomic and functional assignments.

Key Features:

• Gene-Specific Packages (gpkgs): Uses gene-specific packages that can be based on hidden Markov models (HMMs) or DIAMOND databases to identify gene families within metagenomic data.
• Phylogenetic Placement: Places sequences into pre-constructed gene trees to enable phylogenetically informed taxonomic assignment at the family level.
• Read-Level Analysis: Examines each read independently to assess microbial diversity without relying on genome recovery from environmental samples.
• Performance: Benchmarking with in silico and in vitro mock communities demonstrated higher family-level accuracy and processing speeds 2.0 to 3.7 times faster than existing software.
• Support for Marker Genes: Applies specific gpkgs such as 16S rRNA and methyl-coenzyme M reductase (McrA) for targeted taxonomic and functional detection.

Scientific Applications:

• Diversity Analysis: Provides a comprehensive view of microbial community diversity by classifying reads independently in complex metagenomes.
• Functional and Taxonomic Insights: Uses specific gene packages (e.g., 16S rRNA, McrA) to detect taxonomic and functional shifts across environmental gradients, such as depth in wetland metagenomes.
• Novel Sequence Discovery: Enables identification of novel sequences from large databases (e.g., NCBI nr) exemplified by detection of previously uncharacterized phylum-level McrA lineages.

Methodology:

Identifies gene families using gene-specific packages based on HMMs or DIAMOND databases and places identified sequences into pre-constructed gene trees for phylogenetic placement and taxonomic assignment.