search

steineggerlab / conterminator

Detection of incorrectly labeled sequences across kingdoms
GNU General Public License v3.0
79 stars 7 forks source link

GTDB #19

Open andrewjmc opened 3 years ago

andrewjmc commented 3 years ago

Hi,

I would love to clean human contaminated sequences from the GTDB bacteria and archaea (r95) and NCBI viruses and fungi, as classifications are being badly affected in some samples of mine with high human DNA proportion. I already have a concatenated .faa file for kraken, and a seqid2taxid.map file. However, because it is a custom-built database, and incorporates GTDB, the taxids bear no relation to NCBI IDs. I have a names.dmp and nodes.dmp file.

Could I tweak conterminator to process this database? It is a 120 Gb sequence database. I can't see how much RAM is required, but naively following the idea of linear time, I would hope I could process my database in under a day.

Best wishes,

Andrew

martin-steinegger commented 2 years ago

The database module should allow you to download the GTDB database. It will build names.dmp and nodes.dmp based on the GTDB taxonomy.