GTDBTK issues

[ijgretch]

gtdbtk appears to install just fine, but won't work even after re-activating the env. Getting the error: "Gtdbtk does reference not exist or is corrupted." Suggestions?

[bglindner]

This warning is expected -- if you check the step05.sbatch script you'll note that I've provided the path to the database there. You should be good to proceed but if you'd still like to check that gtdbtk has installed successfully, run this first:

GTDBTK_DATA_PATH="/storage/ice-shared/cee6720/01_databases/release214"
export GTDBTK_DATA_PATH

Then you can call gtdbtk -h

[ijgretch]

Okay, yay, I am no longer getting the same error, but now I am getting an "unrecognized argument" error for --mash_db. This argument is in the documentation for classify_wf and is not specified incorrectly in step05.sbatch as far as I can tell. I checked that the path is valid and contains a "gtdb_ref_sketch.msh" as expected, so I am not sure what the problem is.

[bglindner]

Great news!

I've made a quick change to the shared version of GTDB -- which is where this mash database is stored. Can you try this again and let me know if the problem persists? Maybe it was a permissions issue.

[ijgretch]

Hi! Thanks for responding so quickly. I started a new session, reactivated my environment, ran "GTDBTK_DATA_PATH="/storage/ice-shared/cee6720/01_databases/release214" export GTDBTK_DATA_PATH" and then reran the step05 batch script. Same error persisting.

[bglindner]

Would you try editing step05.sbatch and remove the --mash_db flag from the end of the gtdbtk classify_wf command?

So like remove all of this from the command: --mash_db /storage/ice-shared/cee6720/01_databases/release214/mash

If that works, then I'll patch the step05 script in shared/lab03 and encourage others to get a new copy.

[ijgretch]

Alrighty, it gets further along, but there is still an "unexpected" error resulting in an uncontrolled exit. Screenshot of .out:

[bglindner]

Can you run gtdbtk -v and let me know what version of the tool your environment is using?

[ijgretch]

Says it's running 1.0.2 despite having installed 2.3.2 with "micromamba install fastp spades coverm maxbin2 gtdbtk=2.3.2 checkm2 metabat2"

[bglindner]

Try calling micromamba remove gtdbtk and then micromamba install gtdbk=2.3.2

Then launch it again but roll back any changes to step05's script (i.e., so you have the following call to gtdbtk):

gtdbtk classify_wf --batchfile ${manifest} --out_dir ${output} --cpus 12 --mash_db /storage/ice-shared/cee6720/01_databases/release214/mash

[ijgretch]

Ran farther than previously, killed by permission error with guppy. Was messing with tool versions for step04_bakta while this ran. Would that have caused this issue?

[bglindner]

It's possible, see #11 for thoughts on how to proceed!