def _read_sumstats(args, log, fh, alleles=False, dropna=False):
'''Parse summary statistics.'''
log.log('Reading summary statistics from {S} ...'.format(S=fh))
if args.snplist:
log.log(
'and extracting SNPs specified in {S} ...'.format(S=args.snplist))
sumstats = ps.sumstats(fh,
alleles=alleles,
dropna=dropna,
slh=args.snplist)
else:
sumstats = ps.sumstats(fh, alleles=alleles, dropna=dropna)
log_msg = 'Read summary statistics for {N} SNPs.'
log.log(log_msg.format(N=len(sumstats)))
m = len(sumstats)
sumstats = sumstats.drop_duplicates(subset='SNP')
if m > len(sumstats):
log.log('Dropped {M} SNPs with duplicated rs numbers.'.format(
M=m - len(sumstats)))
return sumstats
def _read_sumstats(args, log, fh, alleles=False, dropna=False):
# Parse summary statistics.
log.log('Reading summary statistics from {S} ...'.format(S=fh))
sumstats, n_codings = ps.sumstats(fh, args.additive, args.dominance, args.gxe, alleles=alleles, dropna=dropna)
log_msg = 'Read summary statistics for {N} SNPs.'
log.log(log_msg.format(N=len(sumstats)))
m = len(sumstats)
sumstats = sumstats.drop_duplicates(subset='SNP')
if m > len(sumstats):
log.log('Dropped {M} SNPs with duplicated rs numbers.'.format(M=m - len(sumstats)))
return sumstats, n_codings
def _read_sumstats(args, log, fh, alleles=False, dropna=False):
'''Parse summary statistics.'''
log.log('Reading summary statistics from {S} ...'.format(S=fh))
sumstats = ps.sumstats(fh, alleles=alleles, dropna=dropna)
log_msg = 'Read summary statistics for {N} SNPs.'
log.log(log_msg.format(N=len(sumstats)))
m = len(sumstats)
sumstats = sumstats.drop_duplicates(subset='SNP')
if m > len(sumstats):
log.log(
'Dropped {M} SNPs with duplicated rs numbers.'.format(M=m - len(sumstats)))
return sumstats
