Missing GFF file for SPECTRA_MSD/TB_COVID_GFP

[Tasmin153]

Hi,

I am currently working on phenotype prediction from unique strains of antitubercular agents. Thanks for this great work, SPECTRA would be helpful to test the model that I am working on. I was trying to reproduce the analysis for TB drug splits. I have downloaded the required data files from the dataverse page. But in the "run_baseline.py" file, I find that there are some missing files like input_gff_file, reference_nucleotide, and full_reference_sequence. Could you let me know how can I get these files and run the "run_baseline.py" file correctly?

Input GFF File
reference_nucleotide

[yashaektefaie]

Apologies for the delayed response! You actually do not need those files, these were needed when once upon a time I did data processing per step but I did that once and provide the processed data files. So if you provide None for those entries the script should run if there are more errors though let me know!

[Tasmin153]

Hi, I am trying to run with the following command python run_baseline.py 0 INH logistic_regression 0.1 binary --trial_run True. There are some files missing callbacks.py constants.py, generate_barcode.py, mod_alignment_utils.py, check_mutational_splits.py missing from the SPECTRA_paper/SPECTRA_MSD/TB_Covid_GFP/utils folder. I am guessing most of these methods have been moved to #https://github.com/mims-harvard/SPECTRA/blob/75b59639dcae6adad92af4a34313a75196a2659c/SPECTRA_paper/SPECTRA_MSD/TB_Covid_GFP/utils/general_utility_functions.py. Right now I am commenting out those import statements and importing this file.

However, I could not fine "GenerateBarcode" as mentioned in this line #

SPECTRA/SPECTRA_paper/SPECTRA_MSD/TB_Covid_GFP/dataset/Sequence_Dataset.py

Line 54 in 75b5963

self.fetcher = GenerateBarcode(drug, use_pregenerated)

in any of the utils files or the Sequence_dataset.py file. I have tried commenting it out but it does not work. There are some methods related to barcode in the Sequence_dataset file but I could not find the one matching with this params GenerateBarcode(drug, use_pregenerated)

Traceback (most recent call last):
  File "run_baseline.py", line 395, in <module>
    run_baseline(**params_to_use)
  File "run_baseline.py", line 140, in run_baseline
    sequence_dataset.initialize_encoder()
  File "./SPECTRA/SPECTRA_paper/SPECTRA_MSD/TB_Covid_GFP/run/Sequence_Dataset.py", line 109, in initialize_encoder
    all_train_outputs = [self.__getitem__(i) for i in tqdm(self.return_train_strains(), total=len(self.train_strains))]
  File "./SPECTRA/SPECTRA_paper/SPECTRA_MSD/TB_Covid_GFP/run/Sequence_Dataset.py", line 109, in <listcomp>
    all_train_outputs = [self.__getitem__(i) for i in tqdm(self.return_train_strains(), total=len(self.train_strains))]
  File "./SPECTRA/SPECTRA_paper/SPECTRA_MSD/TB_Covid_GFP/run/Sequence_Dataset.py", line 376, in __getitem__
    sequences = self.get_sequences_barcode(i)
  File "./SPECTRA/SPECTRA_paper/SPECTRA_MSD/TB_Covid_GFP/run/Sequence_Dataset.py", line 370, in get_sequences_barcode
    return self.fetcher.barcode(strain)
AttributeError: 'Sequence_Dataset' object has no attribute 'fetcher'