Issues with adding MAGs to db at the VHM step - number issue?

[Apoorva Prabhu]

Hi,

I ran the iphop add_to_db and predict with test dataset and the pipeline works just fine on our server.

However, I have 2,459 MAGs , and tried to create a db using the below command with iphop_1.3.2, with latest downloaded db.

I first symlinked the original db to my current directory before running the command:

iphop add_to_db --fna_dir /xxxxxx/high_quality_representatives_50%/ --gtdb_dir ../02.gtdbtk/03.combined/ --out_dir results/ --db_dir iphopdir/ -t 30 &

The gtdb_dir contains de novo trees inferred for both bacteria and archaea, as well as the taxonomy table exactly as mentioned in the source page.

The last message I get is that the “[6] Add new genomes to VHM database...
The query file README.txt contains invalid chars, please make sure it is a valid fasta file.”

Highly unlikely but would this be an issue if my files were “.fa” and not “.fna” ? Or is it likely the number of MAGs is an issue? Or would you know how to fix this error please? I have attached the log file.

Thanks!

[Simon Roux]

Hi !

I don’t believe this is a problem of extension or number of MAGs. Would you happen to have a file called “README.txt” in the folder “high_quality_representatives_50%” ? If so, this file should be removed and the process restarted (from the error, it seems like that’s what VHM complains about, i.e. VHM needs to have only fasta files from MAGs in “high_quality_representatives_50%”).

Best,

Simon

[Apoorva Prabhu]

Hi Simon,

Oops! Yes, that’s right - I now symlinked all MAGs to a dir that had no other files except fasta files. The only odd thing I had to do was that the gtdb_dir needs to have “infer” dir with trees and taxonomy, and now the new DB has been created, in under 2 hours.

Thanks so much for the incredible tool, and will run iphop predict now.

Thanks again,

Apoorva

[Simon Roux]

Perfect, glad that was a relatively easy fix, and let me know if you have any other questions !