Example: create and use an LCA database for taxonomic analysis

sourmash-bio/sourmash-examples#2

---

First, download and sketch 64 genomes from Awad et al., 2017 using the instructions in Example: download, sketch, and search a collection of FASTA files. You'll need podar-ref.zip.

Next, you'll need podar-lineage.csv -

curl -L [https://osf.io/4yhjw/download](https://osf.io/4yhjw/download) -o podar-lineage.csv

Now create an LCA database in SQL format:

sourmash lca index podar-lineage.csv podar-ref.lca.sql podar-ref.zip \
    -F sql -C 3 --split-identifiers

Extract one of the Shewanella genomes from podar-ref.zip using sourmash sig grep:

sourmash sig grep OS223 podar-ref.zip -o shew-os223.sig

and now you can classify genomes with lca classify:

sourmash lca classify --query shew-os223.sig --db podar-ref.lca.sql

and you should see it classified correctly:

"NC_011663.1 Shewanella baltica OS223, complete genome",found,Bacteria,Proteobacteria,Gammaproteobacteria,Alteromonadales,Shewanellaceae,Shewanella,Shewanella baltica,Shewanella baltica OS223

You can use lca summarize to classify the genome as if it were a metagenome mixture, too:

sourmash lca summarize --query shew-os223.sig --db podar-ref.lca.sql

and you should see:

50.5%   278   Bacteria;Proteobacteria;Gammaproteobacteria;Alteromonadales;Shewanellaceae;Shewanella;Shewanella baltica;Shewanella baltica OS223   shew-os223.sig:38729c63 NC_011663.1 Shewanella baltica OS223, complete genome
100.0%   550   Bacteria;Proteobacteria;Gammaproteobacteria;Alteromonadales;Shewanellaceae;Shewanella;Shewanella baltica   shew-os223.sig:38729c63 NC_011663.1 Shewanella baltica OS223, complete genome

which indicates that about 50% of the content is not strain specific, and is shared with the other Shewanella in the collection.

Categories

This example belongs to the following categories:

• examples of using taxonomy
• examples using sourmash v4.4 functionality