Cache already aligned genomes

SarahNadeau / burk-nextstrain

An automated nextstrain build for Burkholderia pseudomallei

1 stars 0 forks source link

Cache already aligned genomes #1

Open SarahNadeau opened 1 year ago

SarahNadeau commented 1 year ago

It's expensive to re-make the NCBI database and re-generate the core genome alignment every time a build is updated.

[ ] cache database tables, already aligned assemblies
[ ] query only newly released assemblies
[ ] make sure new assemblies aligned to same reference
[ ] concatenate alignments, update cache
[ ] build tree independently of alignment step

SarahNadeau commented 1 year ago

I forgot that ParSNP requires the core genome to be present in ALL strains, so I can't concatenate alignments to the same reference like I thought. New strains will change the alignment length.

SarahNadeau commented 1 year ago

Use a different alignment method, or call SNPs from SRA data instead?