The EVcouplings has generated a "failed" file

[zhaoyang163SJTU]

Hi, when I use my own data to perform the evcouplings step, relevant files will be generated under the "output" directory, among which the "failed" file will be created:
Traceback (most recent call last):
File "/home/zhaoyang/anaconda3/envs/venusrem/lib/python3.9/site-packages/evcouplings/utils/pipeline.py", line 508, in execute_wrapped
outcfg = execute(**config)
File "/home/zhaoyang/anaconda3/envs/venusrem/lib/python3.9/site-packages/evcouplings/utils/pipeline.py", line 187, in execute
outcfg = runner(**incfg)
File "/home/zhaoyang/anaconda3/envs/venusrem/lib/python3.9/site-packages/evcouplings/compare/protocol.py", line 1638, in run
return PROTOCOLSkwargs["protocol"]
File "/home/zhaoyang/anaconda3/envs/venusrem/lib/python3.9/site-packages/evcouplings/compare/protocol.py", line 1050, in standard
d_intra = intra_dists(
File "/home/zhaoyang/anaconda3/envs/venusrem/lib/python3.9/site-packages/evcouplings/compare/distances.py", line 962, in intra_dists
chain = _prepare_chain(
File "/home/zhaoyang/anaconda3/envs/venusrem/lib/python3.9/site-packages/evcouplings/compare/distances.py", line 851, in _prepare_chain
chain = structures[pdb_id].get_chain(pdb_chain, model)
KeyError: '3da1'

I attempted to use the command "pip install https://github.com/debbiemarkslab/EVcouplings/archive/develop.zip" to change the version of "evcouplings" from 0.2.1 in the "environment.yml" file to 0.3.0. However, it still resulted in an error. Could you please tell me if you know what the problem is?Thanks

[tyang816]

How about trying another protein sequence?

[zhaoyang163SJTU]

Yes, I changed to another sequence. Some of the thresholds resulted in tar.gz files, but the other set of thresholds still encountered the same error.

[tyang816]

Maybe a screenshoot plz?

[zhaoyang163SJTU]

ok

[zhaoyang163SJTU]

Based on the error message, it is speculated that the error might occur during the calculation of structural distances by evcouplings. In venusREM, is it necessary to calculate the structural distances? My understanding is that using evcouplings to obtain comparison files for different thresholds is mainly for generating a2m files. I wonder if this understanding is correct?

[tyang816]

VenusREM doesn't need to calculate the pdb distance, it's a little weird. Different thresholds are used to generate different a2m files and select the most significant one.
Have you ever tried the example sequence?

[zhaoyang163SJTU]

I haven't tried the example sequence. I'll give it a try. If there is a result, I will update it here.

[zhaoyang163SJTU]

Hello, I attempted to comment out the "couplings" and "compare" parameters in the "evcouplings" file. The program was able to generate the tar.gz file normally and no error files were generated, but the "num_significant" column was missing. Therefore, when running the "select_msa.py" script, pandas would encounter an error due to the lack of "num_significant". Thus, commenting out the "couplings" and "compare" parameters is not advisable.
This test use evcouplings = 0.2.1

[zhaoyang163SJTU]

We also attempted to set up the same environment in the hpc (Python 3.9, evcouplings = 0.2.1), and used the fasta files in the example to generate a2m files: A0A140D2T1_ZIKV_Sourisseau_2019 and A0A192B1T2_9HIV1_Haddox_2018_b0.3. However, an error still occurred. Among them, A0A140D2T1_ZIKV_Sourisseau_2019 had a memory overflow due to an excessively long sequence. We will temporarily ignore such errors. In the cases of A0A192B1T2_9HIV1_Haddox_2018_b0.3 and GPO, the program would report the same error.

Attempted to update the bio==1.6.2 in the environment.yml file to bio==1.7.1 as specified in the requirements.txt, but still encountered the same error. Therefore, I don't know the exact cause of the error. Could you please provide the environment file used for generating the a2m file?

[tyang816]

You can find env file in the README. https://github.com/ai4protein/VenusREM/blob/main/environment.yml

[zhaoyang163SJTU]

Hello, I confirm that the environment was set up according to the env file you provided. However, the above error still occurs. I also tried upgrading biopython to version 1.85, but it didn't solve the problem. Do you have any better suggestions?