io.write_plink(problem, out_base_name, verbose=True,
save_node_type=False, save_genotype=True, save_haplotype=False, save_error=False)
# Save statistics and phasing metadata in a separate npz
np.savez(out_base_name + '.stats',
stats=np.array([stats]),
info=np.array([problem.info]),
pedigree=np.array([problem.pedigree]))
plink_cmd_base = '%s --tfile %s' % (bu.PLINK, out_base_name,)
if options.recode:
# First, compute allele frequencies with PLINK
util.run_command('%s --nonfounders --freq --out %s' % (plink_cmd_base, out_base_name))
# Convert frequencies file that to a reference allele recoding
# file (a file containing the list of SNPs and their minor allele letter)
bu.frq_to_minor_file(out_base_name + '.frq', out_base_name + '.mnr')
# Then convert binary PLINK to a recoded 12-recoded TPED, where 1=minor allele for each SNP
out_recoded = out_base_name + '.recoded'
util.run_command('%s --transpose --recode12 --reference-allele %s.mnr --out %s' % \
(plink_cmd_base, out_base_name, out_recoded))
# Reload the recoded problem
for ext in ('nof', 'tped', 'tfam'):
os.rename(out_recoded + '.' + ext, out_base_name + '.' + ext)
genotype = io_genotype.read('plink', 'genotype', tped=out_base_name + '.tped', load_ids=False)
else:
genotype = problem.genotype
# Write problem to file in our (npz)
io.write_npz(problem, out_base_name + '.npz')
num_parts=part_count[options.chrom])
part_names_gxn = bu.partnames(bu.chrnames(options.out_gxn, part=options.chrom),
num_parts=part_count[options.chrom])
(start, stop) = endpoints[options.chrom]
endpoints = util.brange(start, stop, part_size, endpoint=True, dtype=int)
for (part, part_start) in enumerate(endpoints[:-1]):
out = part_names[part]
plink_cmd_base = '%s --cm --bfile %s --chr %d --from-bp %s --to-bp %s --out %s' \
% (bu.PLINK, base_name, options.chrom, part_start, endpoints[part + 1], out)
if options.recode:
# First, compute allele frequencies with PLINK
util.run_command(plink_cmd_base + ' --nonfounders --freq')
# Convert frequencies file that to a reference allele recoding
# file (a file containing the list of SNPs and their minor allele letter)
bu.frq_to_minor_file(out + '.frq', out + '.mnr')
if options.out_gxn:
# Copy FRQ to target output directory
out_frq = part_names_gxn[part] + '.frq'
mkdir_if_not_exists(os.path.dirname(out_frq))
shutil.copy(out + '.frq', out_frq)
# Then convert binary PLINK to a 12-recoded TPED, where 1=minor allele for each SNP
cmd = '%s --transpose --recode12 --reference-allele %s.mnr' % (plink_cmd_base, out)
util.run_command(cmd)
else:
# No recoding, just convert binary to 2-recoded TPED. PLINK assigns "1" to
# the first-encountered allele in the file for each SNP.
cmd = '%s --transpose --recode12' % (plink_cmd_base,)
util.run_command(cmd)
