singscore_example.py

import os
from singscore.singscore import *

BASE_DIR = os.path.dirname(__file__)

entrez_sample_path = os.path.normpath('{}/{}'.format(BASE_DIR,
                                'singscore/test/test_data/entrez_sample.txt'))

sig_path = os.path.normpath('{}/{}'.format(BASE_DIR,
                                'singscore/test/test_sigs/tgfb_upDown.txt'))


data = pandas.read_csv(open(entrez_sample_path, 'r'), header=
'infer', sep='\t')
data = data.set_index(keys='EntrezID')

# prepare signatures
sigs = pandas.read_csv(open(sig_path, 'r'), header=
'infer', sep='\t')
# subset the data for up and down
up = sigs[sigs['upDown'] == 'up']
down = sigs[sigs['upDown'] == 'down']

# get a list of ids
up = list(up['EntrezID'])
down = list(down['EntrezID'])

# scoring
scored_data = score(up_gene=up, down_gene=down, sample=data,
                    norm_method='theoretical', full_data=True)
# print(scored_data)



# plot dispersion of scores
output_path = os.path.normpath('{}/{}'.format(BASE_DIR,
                                    'singscore/test/output/dispersion.png'))

plotdispersion(scored_data, ctrlstring='Ctrl', teststring='TGFb',
               testlabel='TGFb', colour_1='g', colour_2='b',
               outpath=output_path, show=False)


# plotting distribution of ranks single sample, D_Ctrl_R1 -  first find ranks
ranked_data = rank(up_gene=up, down_gene=down, sample=data[['D_Ctrl_R1']],
                   norm_method='theoretical')
# print(ranked_data)
# plot
output_path = os.path.normpath('{}/{}'.format(BASE_DIR,
                                    'singscore/test/output/barcode.png'))

x = plotrankdist(ranks=ranked_data, colour_1='r', colour_2='b',
                 output=output_path, show=False)
# for significance test
output_path = os.path.normpath('{}/{}'.format(BASE_DIR,
                                        'singscore/test/output/nulldist.png'))
permd = permutate(data, n_up=193, n_down=108)
pvals = empiricalpval(permutations=permd, score=scored_data)
# print(pvals)
nulldistribution(permutations=permd, score=scored_data, nrows=2, ncols=5,
                 threshold=0.05, outpath=output_path, show=False)