AlphaFold Multimer about gget

[lauraluebbert]

Hi Esa-Pekka, gget alphafold already supports multimers. Simply pass multiple sequences as a list, or a fasta file containing more than one sequence, and it will automatically use the multimer model. Let me know if it works for you! :)

[epkumpu]

Ahhh, my bad. I was using the : separator as in the colabfold implementation which obviously didn’t work. Thank you for the response! EP

[epkumpu]

Hi again!

I have an issue for follow-up: I ran the prediction with one sequence and it worked alright, but when I tried with two sequences (now correctly input as a .fasta file!), I get a "No space left on device" error. I have set this up to run on a cluster in an area where I have plenty of quota left, so my guess is it is trying to write temporary files into another location.

Is there a way to set the location for any temporary files that gget is writing manually?

Here is the error log:

openmm  7.5.1  environment loaded 
2022-09-12 15:26:40.813182: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is
 optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in perfo
rmance-critical operations:  AVX2 AVX512F AVX512_VNNI FMA
To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
2022-09-12 15:26:44.994756: I tensorflow/core/util/util.cc:169] oneDNN custom operations are on. You may s
ee slightly different numerical results due to floating-point round-off errors from different computation 
orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
/scratch/project_2004941/gget/lib/python3.9/site-packages/haiku/_src/data_structures.py:37: FutureWarning:
 jax.tree_structure is deprecated, and will be removed in a future release. Use jax.tree_util.tree_structu
re instead.
  PyTreeDef = type(jax.tree_structure(None))
Mon Sep 12 15:28:46 2022 INFO Validating input sequence(s).
Mon Sep 12 15:28:46 2022 INFO Finding closest source for reference database.
Using the multimer model with 2 sequences.
Traceback (most recent call last):
  File "/opt/conda/envs/env/lib/python3.9/urllib/request.py", line 272, in urlretrieve
    tfp.write(block)
  File "/opt/conda/envs/env/lib/python3.9/tempfile.py", line 478, in func_wrapper
    return func(*args, **kwargs)
OSError: [Errno 28] No space left on device

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/scratch/project_2004941/gget/bin/gget", line 8, in <module>
    sys.exit(main())
  File "/scratch/project_2004941/gget/lib/python3.9/site-packages/gget/main.py", line 1518, in main
    alphafold(
  File "/scratch/project_2004941/gget/lib/python3.9/site-packages/gget/gget_alphafold.py", line 404, in al
phafold
    source = f.result()
  File "/opt/conda/envs/env/lib/python3.9/concurrent/futures/_base.py", line 439, in result
    return self.__get_result()
  File "/opt/conda/envs/env/lib/python3.9/concurrent/futures/_base.py", line 391, in __get_result
    raise self._exception
  File "/opt/conda/envs/env/lib/python3.9/concurrent/futures/thread.py", line 58, in run
    result = self.fn(*self.args, **self.kwargs)
  File "/scratch/project_2004941/gget/lib/python3.9/site-packages/gget/gget_alphafold.py", line 116, in fe
tch
    request.urlretrieve(test_url_pattern.format(source))
  File "/opt/conda/envs/env/lib/python3.9/urllib/request.py", line 275, in urlretrieve
    reporthook(blocknum, bs, size)
  File "/opt/conda/envs/env/lib/python3.9/tempfile.py", line 496, in __exit__
    result = self.file.__exit__(exc, value, tb)
OSError: [Errno 28] No space left on device

[lauraluebbert]

gget will currently create a "tmp" folder in your home directory ("~/tmp/jackhmmer/") for the Jackhmmer search. I think adding an option to change this path is a great idea for a future version. The temporary files will take up to ~2 GB (in case it is possible to free this space until I have implemented your request).