run_EGF_HER2_analysis.py

#Import preamble of kstar and other necessary functions
import pandas as pd
import os
import pickle

from kstar import config, helpers, calculate


import conf #applications config for supplementary directory

if __name__ ==  '__main__': #added this to prevent recursive spooling of multiprocessing

	data_dir = conf.SUPPLEMENTS_DIR+'Control_experiments/EGF_184A1_HER2_WolfYadlin2006/'

	odir = './data/predictions/'
	expName = 'EGF_HER2'

	experiment = pd.read_csv(f"{data_dir}/MAPPED/{expName}_mapped.tsv", sep='\t')

	networks = {}
	networks['Y'] = pickle.load(open(config.NETWORK_Y_PICKLE, "rb" ) )


	if not os.path.exists(f"{odir}/{expName}"): 
	    os.mkdir(f"{odir}/{expName}")
	activity_log = helpers.get_logger(f"activity_{expName}", f"{odir}/{expName}/activity_{expName}.log")

	data_columns = None
	agg = 'mean' # if a non-NaN value appears at all, use it
	threshold = 0.8 #require they are within 80% of what was seen at the normalized 4' mark
	#Save the SLIM versions at each step, in case of failure/loss of kernel, etc.

	kinact_dict = calculate.run_kstar_analysis(experiment, activity_log, networks, phospho_types = ['Y'], data_columns = data_columns, agg =agg, threshold = threshold,  greater = True)
	calculate.save_kstar_slim(kinact_dict, expName, f"{odir}/{expName}")


	num_random_experiments=150
	target_alpha=0.05
	calculate.normalize_analysis(kinact_dict, activity_log, num_random_experiments, target_alpha)
	calculate.save_kstar_slim(kinact_dict, expName, f"{odir}/{expName}")


	kinact_dict['Y'].calculate_Mann_Whitney_activities_sig(activity_log, number_sig_trials = 100)
	calculate.save_kstar_slim(kinact_dict, expName, f"{odir}/{expName}")