PI score calculation failed

[jsquaredosquared]

In the statistics.csv file, the following fields are empty for all predictions:

Num_intf_residues	Polar	Hydrophobic	Charged	contact_pairs	sc	hb	sb	int_solv_en	int_area	pi_score

Why are these fields empty?

Looking at the log file for the compute_stats rule, I get the following error:

Building DAG of jobs...
Using shell: /usr/bin/bash
Provided cores: 1 (use --cores to define parallelism)
Rules claiming more threads will be scaled down.
Provided resources: mem_mb=8000, mem_mib=7630, disk_mb=1000, disk_mib=954, walltime=1440, attempt=1
Conda environments: ignored
Select jobs to execute...

[Mon Aug 26 01:16:18 2024]
rule compute_stats:
    input: /scr/user/jeffrey.o/alphapulldown_output/predictions/Q6BDS2_P51149/completed_fold.txt
    output: /scr/user/jeffrey.o/alphapulldown_output/predictions/Q6BDS2_P51149/statistics.csv
    jobid: 0
    reason: Forced execution
    wildcards: fold=Q6BDS2_P51149
    resources: mem_mb=8000, mem_mib=7630, disk_mb=1000, disk_mib=954, tmpdir=/tmp, walltime=1440, attempt=1, partition=cpu-epyc-genoa,cpu-epyc

Activating singularity image /home/user/jeffrey.o/AlphaPulldownSnakemake/.snakemake/singularity/3f20617ccba864758b2a437ef2fde35c.simg
INFO:    Could not find any nv files on this host!
INFO:    Converting SIF file to temporary sandbox...
I0826 01:18:23.496333 22361792987968 get_good_inter_pae.py:120] now processing Q6BDS2_P51149
E0826 01:18:31.441169 22361792987968 pdb_analyser.py:173] PI score calculation failed: Command '['/bin/bash', '-c', 'source activate pi_score && export PYTHONPATH=/software:$PYTHONPATH && python /software/pi_score/run_piscore_wc.py -p /tmp/tmp.ubjPZQDmEl/Q6BDS2_P51149/ranked_0.pdb -o /tmp/tmp.ubjPZQDmEl/pi_score_outputs/Q6BDS2_P51149 -s 2 -ps 10']' returned non-zero exit status 1.. Command: ['/bin/bash', '-c', 'source activate pi_score && export PYTHONPATH=/software:$PYTHONPATH && python /software/pi_score/run_piscore_wc.py -p /tmp/tmp.ubjPZQDmEl/Q6BDS2_P51149/ranked_0.pdb -o /tmp/tmp.ubjPZQDmEl/pi_score_outputs/Q6BDS2_P51149 -s 2 -ps 10']. Error: Last system error message: No such file or directory
 SC:   Open failed: File: /tmp/root/sc_tmp.19787                                                                                     

W0826 01:18:31.442494 22361792987968 pdb_analyser.py:192] Skipping interface due to unexpected format: None
E0826 01:18:31.442568 22361792987968 pdb_analyser.py:319] An unexpected error occurred while merging the PI score data: local variable 'pi_score_files' referenced before assignment
W0826 01:18:31.442638 22361792987968 pdb_analyser.py:192] Skipping interface due to unexpected format: None
I0826 01:18:34.871306 22361792987968 get_good_inter_pae.py:168] done for Q6BDS2_P51149 1 out of 1 finished.
INFO:    Cleaning up image...
[Mon Aug 26 01:18:56 2024]
Finished job 0.
1 of 1 steps (100%) done

What can be done to fix this?

[DimaMolod]

Hi @jsquaredosquared
it could be that pi-score fails because no interface has been found.
Could you enter your container:
singularity shell /home/user/jeffrey.o/AlphaPulldownSnakemake/.snakemake/singularity/3f20617ccba864758b2a437ef2fde35c.simg
and try running this command directly:
source activate pi_score && \ export PYTHONPATH=/software:$PYTHONPATH && \ python /software/pi_score/run_piscore_wc.py \ -p /tmp/tmp.ubjPZQDmEl/Q6BDS2_P51149/ranked_0.pdb \ -o /tmp/tmp.ubjPZQDmEl/pi_score_outputs/Q6BDS2_P51149 \ -s 2 \ -ps 10
then you should get a clearer error message.
Also, please open your Q6BDS2_P51149 model, e.g., in PyMol, and check whether there is an interface.
Please let me know about your findings.
Cheers
d

[jsquaredosquared]

All right. I will patiently await your fix :D

[DimaMolod]

Hi @jsquaredosquared and I am sorry for the delay.
Please feel free to update the docker images and run again, the problem is fixed now.
I also have less positive news: we are going to remove ccp4 package from the analysis pipeline image very soon due to some legal issues. But we will also update the documentation and explain how AP users could build the image themselves and keep using it with our snakemake pipeline.

[jsquaredosquared]

I am sorry for the delay.

No worries. Thanks for your hard work :)

we are going to remove ccp4 package

Sucks that you have to remove ccp4. I hope it won't be too difficult to incorporate it.

How helpful is it? Is it an absolutely necessary component, or can AlphaPulldown work well without it?

[DimaMolod]

No, it's not necessary at all. Without ccp4 it will all work up until the issue you reported here - these columns in the table will be always empty:

Num_intf_residues	Polar	Hydrophobic	Charged	contact_pairs	sc	hb	sb	int_solv_en	int_area	pi_score

I hope it will be easy enough to install ccp4 within the container and run AP as before.
Thanks again for your report and happy modeling :-)

[jsquaredosquared]

Hi @DimaMolod ,

Sorry to bother you again. This time, for most of the predictions, the columns were filled in except for the pi_score, which is completely missing in the output file :/