Human readable files/folders as Genus_species_strain

In my (older unreleased scruffier) version of this tool I create a hierarchy like this:

Kingdom/
   Genus/
       species/
            strain/
                    blah.gbk
                    blah.fna
                    blah.gff

I currently set blah to Genus_species_strain but that loses the GCA_xxxx accession. I was thinking of having a a ‘mirror’ folder of symlinks with human readable names.

It was tricky to extract the strain as it appears in up to 3 different columns sometimes, but it mostly works.

The reason for this is to make it easy to work with sequences and get human readable labels etc.

[kblin]

So, apart from making up my mind how to best do the cleanup, I’ve got a working implementation of this. Expect this in the next version of ncbi-genome-download.

[tseemann]

FIrst I try and match the strain from column 8 $x[7] =~ m/^(\S+)\s+(\S+)(\s+.*)?$/

Then if that fails I try column 9: $x[8] =~ m/^strain=(.*)$/)

If that fails I use the assembly ID: $strain ||= $x[0];

And then i sanitize the string:

$s =~ s/[\[\]]/ /g; $s =~ s/^\s+//g; $s =~ s/\s+$//g; $s =~ s/\s+/_/g; $s =~ s/['"]/_/g; $s =~ s{[/;:]}{-}g; $s =~ s/_+/_/g;