def build_plot_by_profile(filename, profile, oligo_id):
rep_reads = {}
fetchReads(filename, rep_reads, oligo_id)
setFigType('png')
fig = plotProfiles([profile], [rep_reads], [43], [False], ['Predicted'],
title='In Frame: %.1f%%' % rep_reads[FRAME_SHIFT])
return fig
def plot_predictions(theta_file, target_seq, pam_idx, out_filename=None):
if pam_idx < 0 or pam_idx >= (len(target_seq) - 3):
raise Exception('PAM idx out of range')
if sum([x in ['A', 'T', 'G', 'C'] for x in target_seq]) != len(target_seq):
raise Exception('Sequence must be composed of A,T,G,or C only')
if len(target_seq) < 20 or pam_idx < 13 or pam_idx > len(target_seq) - 7:
raise Exception(
'Sequence too short or PAM too close to edge of sequence (must have at least 10nt either side of cut site)'
)
if target_seq[pam_idx + 1:pam_idx + 3] != 'GG':
raise Exception('Non NGG PAM (check correct index of PAM)')
profile, rep_reads, in_frame = predictMutations(theta_file, target_seq,
pam_idx)
if not out_filename:
out_filename = '%s_%d.txt' % (target_seq, pam_idx)
fout = io.open(out_filename, 'w')
fout.write(u'@@@%s\n' % ('%.1f' % in_frame))
writePredictedProfileToSummary(profile, fout)
fout.close()
setFigType('png')
fig = plotProfiles([profile], [rep_reads], [pam_idx], [False],
['Predicted'],
title='In Frame: %.1f%%' % in_frame)
return fig
compileGenIndelReads(gen_indel_dir=old_gen_dir,
out_dir=reads_dir,
sample_dirs=old_dirs)
setReadsDir(reads_dir)
#Compute features for each indel
features_dir = getHighDataDir() + '/features_for_gen_indels'
computeFeaturesForGenIndels(gen_indel_dir=new_gen_dir,
out_dir=features_dir)
computeFeaturesForGenIndels(gen_indel_dir=old_gen_dir,
out_dir=features_dir)
setFeaturesDir(features_dir)
if __name__ == '__main__':
setIndelGenExe('/usr/local/bin/indelgen')
setPlotDir('/results/plots')
setFigType('png')
shutil.copytree('/data/predicted_vs_measured_example',
'/results/predicted_vs_measured_example')
prepareExample('/results/predicted_vs_measured_example')
#Predict mutations using pre-trained model and compare to actual (for one oligo only)
theta_file = getHighDataDir(
) + '/model_output_10000_0.01000000_0.01000000_-0.607_theta.txt_cf0.txt'
computeAndComparePredicted(theta_file,
selected_id='Oligo35785',
out_dir='.')
def build_plot_by_profile(filename, profile, oligo_id):
rep_reads = {}
fetchReads(filename, rep_reads, oligo_id)
setFigType('png')
fig = plotProfiles([profile], [rep_reads], [43], [False], ['Predicted'])
return fig
