search

baker-laboratory / RoseTTAFold-All-Atom

Other
613 stars 108 forks source link

Issues with Database Access and File Handling in RosettaFold All-Atom Pipeline #104

Open Javaaaad opened 4 months ago

Javaaaad commented 4 months ago

Hello, I am experiencing multiple errors when running the RosettaFold all-atom inference pipeline, specifically related to database accessibility and file handling. Here's a detailed description of the issues encountered during execution:

  1. Database Access Errors:

    • The system reports "ERROR: Could find neither hhm_db nor a3m_db!" consistently across multiple runs with different E-value cutoffs. This suggests a potential issue in accessing or recognizing the HHblits databases.

  2. File Handling Issues:

    • Several crucial files expected by the script are reported as not found, e.g., t000_.1e-10.a3m, t000_.msa0.a3m, which prevents the pipeline from progressing beyond the sequence alignment stage.

  3. Error Messages and Warnings:

    • Various error messages indicate that files expected as output from one stage are unavailable for subsequent stages, implying either the files are not being created or are being incorrectly referenced.
    • Additionally, a warning about the missing _self_ keyword in protein_sm.yaml suggests potential issues with configuration handling.

Steps to Reproduce:

  1. Run the inference script with the following command: python -m rf2aa.run_inference --config-name protein_sm.yaml hydra.searchpath=[file:///path/to/RoseTTAFold-All-Atom/rf2aa/config/inference]

  2. Errors and warnings appear during the execution related to HHblits and subsequent data processing steps.

Please see attached for the full log.

RFAA_log.txt

Thank you for looking into this.