def writeMatlabScript(sample, species, tempDir, lowessDir, segmentDir):
segVars = cfg.Segment()
matlabName = ''.join(sample.split('_'))
matlabName = ''.join(matlabName.split('-'))
matlabName = ''.join(matlabName.split('.'))
scriptFile = tempDir + matlabName + '.m'
OUT = open(scriptFile, 'w')
OUT.write('%Sample specific variable definitions\n')
OUT.write(str("refFile = '" + segVars.binDict[species] + "';\n"))
OUT.write(str("binFile = '" + lowessDir + sample + ".lowess.txt';\n"))
OUT.write(str("saveFile = '" + segmentDir + sample + ".segments.txt';\n"))
# OUT.write(str("chromNum = " + str(segVars.chromNumDict[species]) + ";\n"))
OUT.write(str("alpha = " + str(segVars.CBSalpha) + ";\n"))
OUT.write('\n\n\n\n\n%Generic processing code\n')
IN = open(segVars.matlabBase, 'r')
for x in IN:
OUT.write(x)
OUT.close()
IN.close()
return matlabName
def runNormalizeOne(species, infile, methodRef, outfile):
normVars = cfg.Segment()
binArray = common.importInfoFile(normVars.binDict[species], [0, 2, 6],
'normref',
skiprows=1)
data = np.loadtxt(infile, usecols=[3], dtype='int')
lowessData = runLowess(data, binArray['gc'])
if methodRef:
lowessData = lowessData - methodRef
np.savetxt(outfile, lowessData)
def runMakeMethodRef(species, sampleList, methodName, lowessDir):
if len(sampleList) < 10:
return False
normVars = cfg.Segment()
binArray = common.importInfoFile(normVars.binDict[species], [0, 2, 6],
'normref',
skiprows=1)
xLocs = [x for x, y in enumerate(binArray['chrom']) if y == 'chrX']
yLocs = [x for x, y in enumerate(binArray['chrom']) if y == 'chrY']
mergeArray = np.zeros(len(binArray), dtype='int')
sampleCount = 0
for i in sampleList:
data = np.loadtxt(i, usecols=[3], dtype='int')
if sum(data) < 500000:
continue
mergeArray = mergeArray + data
sampleCount += 1
if sampleCount < 10:
return False
mergeArray, maleTest = adjustSexChroms(mergeArray, xLocs, yLocs,
sampleCount)
lowessData = runLowess(mergeArray, binArray['gc'])
if not maleTest:
lowessData[yLocs] = len(yLocs) * [0.]
np.savetxt(lowessDir + methodName + '.methodRef.lowess.txt', lowessData)
printText = '\tAmplification method reference for ' + methodName + ' has been generated from ' + str(
sampleCount) + ' samples '
if maleTest:
printText += 'which included at least one male\n'
else:
printText += 'with included no male samples\n'
print(printText)
return lowessData
#!/usr/bin/python import sys import os import inspect import numpy as np currentdir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe()))) parentdir = os.path.dirname(currentdir) sys.path.insert(0,parentdir) import common import config as cfg normVars = cfg.Segment() from statsmodels.nonparametric.smoothers_lowess import lowess as lowess #Lowess regression to correct bin counts for gc-content bias# def runLowess(counts, gc): counts = counts + 1 counts = counts / np.median(counts) counts = np.log(counts) lowessModel = lowess(counts, gc, frac=0.05, return_sorted=False)