Train own model using own dataset
Opened this issue · 0 comments
We have generated some screening data from some in vitro experiments. Altogether we have 196 target sequences and single-end sequencing gave us one FASTQ file.
I am trying to use the files under sample_training_codes/
to train a model, but ran into these two problems:
-
Since I have multiple target sequences and corresponding references, how should I config the file
sample.ini
? -
I added these lines
with open("sample_mutation_pattern_dict.pickle", "wb") as o:
pickle.dump(m_pattern_dict, o)
with open("sample_mutation_rate_dict.pickle", "wb") as o:
pickle.dump(m_rate_dict, o)
to sample_training_codes/main.py
so that two pickle files (sample_mutation_pattern_dict.pickle
,
sample_mutation_rate_dict.pickle
) were saved. But while using sample_models/pickles/pickle_to_csv.py
to convert the two pickle
files to csv
file, the file is required: "/Users/hideto/Downloads/pickles/EGFP_mutation_rate_dict.pickle"
which is not included in this repo.