Template all atom mask was all zeros error

[double-vin]

I also encountered this problem. https://github.com/hpcaitech/FastFold/issues/106

I get the following error message: ` running in multimer mode... Traceback (most recent call last): File "/public/home/FastFold/lib/conda/envs/fastfold/lib/python3.8/site-packages/fastfold/data/templates.py", line 859, in _process_single_hit features, realign_warning = _extract_template_features( File "/public/home/FastFold/lib/conda/envs/fastfold/lib/python3.8/site-packages/fastfold/data/templates.py", line 651, in _extract_template_features raise TemplateAtomMaskAllZerosError( fastfold.data.templates.TemplateAtomMaskAllZerosError: Template all atom mask was all zeros: 6qx9_A1. Residue range: 49-100

During handling of the above exception, another exception occurred:

Traceback (most recent call last): File "inference.py", line 546, in main(args) File "inference.py", line 165, in main inference_multimer_model(args) File "inference.py", line 280, in inference_multimer_model feature_dict = data_processor.process_fasta( File "/public/home/FastFold/lib/conda/envs/fastfold/lib/python3.8/site-packages/fastfold/data/data_pipeline.py", line 1165, in process_fasta chain_features = self._process_single_chain( File "/public/home/FastFold/lib/conda/envs/fastfold/lib/python3.8/site-packages/fastfold/data/data_pipeline.py", line 1114, in _process_single_chain chain_features = self._monomer_data_pipeline.process_fasta( File "/public/home/FastFold/lib/conda/envs/fastfold/lib/python3.8/site-packages/fastfold/data/data_pipeline.py", line 942, in process_fasta template_features = make_template_features( File "/public/home/FastFold/lib/conda/envs/fastfold/lib/python3.8/site-packages/fastfold/data/data_pipeline.py", line 76, in make_template_features templates_result = template_featurizer.get_templates( File "/public/home/FastFold/lib/conda/envs/fastfold/lib/python3.8/site-packages/fastfold/data/templates.py", line 1167, in get_templates result = _process_single_hit( File "/public/home/FastFold/lib/conda/envs/fastfold/lib/python3.8/site-packages/fastfold/data/templates.py", line 888, in _process_singlehit "%s%s (sum_probs: %.2f, rank: %d): feature extracting errors: " TypeError: must be real number, not NoneType `

this here is the input file: input.txt

In the fastfold/data/templates.py file，I modified the 884th line warning=None to solve this problem.

[Gy-Lu]

Hi, could you share your fasta file? I'll try to reproduce it, In my experience, the main reason of this problem is the illegal sequence in the fasta file.

[double-vin]

here is the input file: input.zip

[double-vin]

This input file works normally on alphafold2. What's the problem？I look forward to your reply.

[Gy-Lu]

Hi, sorry for being so late. I have reproduced the error, and I am solving it. This error would be fixed this weekend.

[Gy-Lu]

Hi, this bug has been fixed. If it is not able to solve your problem, you can reopen this issue.

[double-vin]

Thank you! My problem is solved.