example.ncbi-assemblies+protein

Building a sourmash database from NCBI assemblies

This directory contains a Makefile that will build a sourmash database for the genome in GCF_000018865.1_ASM1886v1_genomic.fna, as well as construct a proteome using prodigal and build a protein signature.

Run make to run the entire pipeline. (You'll need sourmash v4.4.0 installed.)

The Makefile does the following:

1. Produce amino acid files containing the genes.

The Makefile will first run the snakemake workflow in Snakefile to build _protein.faa files for all the .fna files in the current directory using prodigal.

2. Build a summary of the genome and proteome files in the 'fromfile' format

Next, the Makefile uses the script ../genbank-to-fromfile.py to scan the genomes and produce a summary file, build.csv, that contains names and source genomes for sourmash signatures. (In this case, there's only one genome, note!)

Names for the genomes are taken from the NCBI assembly_summary.txt file that is distributed with Refseq and Genbank assemblies.

3. Create the signatures using sourmash sketch

Finally, the Makefile runs

sourmash sketch fromfile build.csv -p dna -p protein -o all.zip

to sketch all of the genomes in build.csv. The parameter string -p dna tells sourmash to construct DNA sketches using the default parameters, and -p protein tells sourmash to construct protein sketches, too.

The names for the output signatures are taken from build.csv.

Finished!

You can run sourmash sig summarize all.zip to get a summary of the contents of the zip file, or sourmash sig describe all.zip to get a listing of all the signatures.