Question about some file path

[zw-SIMM]

Nice work and beautiful codes! But some question about file path. What is the file alphafold_enzymes.p and uniprot2sequence_standard_set_structs.p in clipzyme/datasets/enzyme_screening.py ? How to generate these files?

     self.alphafold_files = pickle.load(open("/home/datasets/alphafold_enzymes.p", "rb"))

....

    @staticmethod
    def set_args(args) -> None:
        args.dataset_file_path = (
            "/home/datasets/uniprot2sequence_standard_set_structs.p"
        )

Additionally, args.load_wln_cache_in_dataset seems missing.

[pgmikhael]

Hi,

Thank you for your interest!

Here,

• alphafold_enzymes.p would be a set of protein (UniProt) IDs that have an alpha-fold structure saved locally (this is only used in the skip_sample method to save time from checking whether the path exists with something like os.path.exists)
• uniprot2sequence_standard_set_structs.p: is a dict mapping a protein ID to its sequence, similar to the uniprot2sequence.p file.

Thanks for pointing out the --load_wln_cache_in_dataset flag. This should be False / not used. We will fix it in the next code update as well!

[zw-SIMM]

Hi,

Thank you for your interest!

Here,

• alphafold_enzymes.p would be a set of protein (UniProt) IDs that have an alpha-fold structure saved locally (this is only used in the method to save time from checking whether the path exists with something like skip_sample``os.path.exists)
• uniprot2sequence_standard_set_structs.p: is a dict mapping a protein ID to its sequence, similar to the file.uniprot2sequence.p

Thanks for pointing out the flag. This should be False / not used. We will fix it in the next code update as well!--load_wln_cache_in_dataset

Many thanks for your reply! The simplest method seems to comment them in load_dataset method and skip_samplemethod ?

[pgmikhael]

Yes I think that should work. I'll also just clean up because some code is unused in the final version.