agil2ch

gistfile1.txt

from preproc.pipeline import Pipeline
from preproc.pipeline.heplers import extract_supplementary_files, match_samples_in_dir
from ...task import TaskDescription
from os.path import join, basename
from rpy2.robjects import packages as rpackages
import numpy as np
import pandas as pd
import re
from glob import glob


class Agilent2ChPipeline(Pipeline):
    def run(self, task_description: TaskDescription, in_dir: str, out_dir: str):
        suppl_files = [join(in_dir, basename(f)) for f in task_description.supplementary_files]
        extract_supplementary_files(suppl_files, in_dir)
        matched_samples = match_samples_in_dir_ch2(in_dir, [".txt.gz"], task_description.samples)

        # [(file_name, accession), (file_name, accession)]
        for file_name, accession in matched_samples:
            data = process(file_name)

            series_to_csv_compressed(data['channel1'], out_file)
            series_to_csv_compressed(data['channel2'], out_file)

        sample_names = glob(join(in_dir, '*.txt.gz'))

        scan_upc = rpackages.importr('SCAN.UPC')
        biobase = rpackages.importr('Biobase')


        for sample_name in sample_names:
            res = scan_upc.SCAN_TwoColor(sample_name)
            exprs = pd.DataFrame(data=np.log2(np.array(biobase.exprs(res))), columns=['channel1', 'channel2'])

            cleaned_sample_name = re.sub(r'(GSM\d+).*', '\1', sample_name)

            exprs['channel1'].to_csv(join(out_dir, cleaned_sample_name+'_1'), header=None)
            exprs['channel2'].to_csv(join(out_dir, cleaned_sample_name+'_2'), header=None)