2024-05-17 update: This tool no longer works as the database reference is not being served anymore.

A command line tool to quickly identify closely related genomes using a marker-gene based approach.

~500,000 RefSeq/GenBank (RefBank) and ~280,000 Unified Human Gastrointestinal Genome (UHGG) were downloaded and analyzed to identify phylophlan marker genes. These marker genes are stored in a database and all mapped back to the original genome. Given a query genome, GenomeSearch extracts marker genes from the query genome and searches against the selected database to quickly identify closely related genomes. This can also be run in meta mode where each contig in the FASTA file is treated independently.

This tool has advantages to other approaches, such as kmer-based classification systems, because it doesn't need to load a massive index into memory, and it doesn't depend on external taxonomic classifications. It's pretty fast, too!

Just type in

pip install genomesearch

And then

genomesearch install

It will then prompt you to enter a number between 1 and 400, representing the number of marker gene databases you want to have available. If you plan on using GenomeSearch in meta mode, I recommend 400. If you plan on using it only to analyze isolates, the default of 150 should suffice.

That's it, you are ready to go!

If you want to search a single isolate genome against the RefBank database, type:

genomesearch isolate refbank <myIsolate.fna>

If you want to search against the UHGG database, which is a detailed database of Human Microbiome species, use:

genomesearch isolate uhgg <myIsolate.fna>

If you want to perform a metagenomic search against refbank, use:

genomesearch meta refbank <myMetagenome.fna>

And if you want to perform a metagenomic search against uhgg, use:

genomesearch meta uhgg <myMetagenome.fna>

This program is unpublished and has not been peer-reviewed. If you wish to cite the program, please submit an issue and we will help you.