Attribution error while predicting chromatin profiles with fasta file

[pyareedash]

Hi,
While trying to run the following command I can't get past this error, my hunch is it's from selene-sdk (newest version 0.5.0)

sh 1_sequence_prediction.sh <input-file> <genome> <output-dir> --cuda

And I get,

use_cuda: True Traceback (most recent call last): File "1_sequence_prediction.py", line 73, in <module> parse_configs_and_run(configs) ... if not self.reference_sequence._initialized: AttributeError: type object 'Genome' has no attribute '_initialized'

Thank you in advance!

[kathyxchen]

Hi, thanks for your message. How did you install selene? If you can share your exact command that would also help. Happy to discuss here or via email chen.kathleenm@gmail.com if that's easier.

[pyareedash]

Hi,
Thank you for your response!
I tried installing with conda/mamba for e.g., mamba install -c bioconda selene-sdk and mamba install selene-sdk=0.5.
Shall I try installing it from the GitHub repo maybe?

[pyareedash]

Update: I tried pip install selene-sdk==0.5.1 and seems to work (running with no errors).

I will update more once it completes the run :)

[kathyxchen]

Noted thanks for the update and keep me posted! I'll make a note to check if we need to be more clear on versioning, conda installation details, etc

[pyareedash]

Yes, working with selene-sdk==0.5.1 (via pip installation)
Thank you very much!