convert pos and samples from replaced files

convertSNP.py

#!/usr/bin/env python
#=======================================================
# convert pos and samples
# row name is sample name, and column name is position
# value is the number of alt (0,1,2)
#=======================================================

# sampledict: {'sample1':['0','2',...], ...}
def convertSNP(infile, outfile):
    with open(infile) as f:
        header = f.readline().strip().split('\t')
        samplenames = header[6:]
        sampledict = {}
        poslist = []
        #print samplenames

        for line in f:
            pos = line.strip().split('\t')[1]
            poslist.append(pos)
            sampleinfos = line.strip().split('\t')[6:]

            for samplename, snp in zip(samplenames,sampleinfos):
                if samplename not in sampledict:
                    sampledict[samplename] = [snp]
                else:
                    sampledict[samplename].append(snp)
                    
    with open(outfile, 'w') as out:
        out.write('Sample\t'+'\t'.join(poslist)+'\n')
        for samplename, snp in sampledict.items():
            out.write(samplename+'\t'+'\t'.join(snp)+'\n')
          

if __name__ == '__main__':
    import sys

    if len(sys.argv) < 3:
        print "Usage: python %s <infile> <outfile>" % sys.argv[0]
        exit(1)
    infile, outfile = sys.argv[1:3]

    convertSNP(infile, outfile)