def loadgensnpy(filename, source = None):
'''
Snpy wrapper, optional source as named param.
source = one of ``"23andme"``, ``"vcf"``,
``"decodeme"`` or ``"ftdna"``;
'''
snpy_generator = sn.parse(filename, source)
return snpy_generator
def loadgensnpy(filename, source = None):
'''
Snpy wrapper, optional source as named param.
source = one of ``"23andme"``, ``"vcf"``,
``"decodeme"`` or ``"ftdna"``;
'''
snpy_generator = sn.parse(filename, source)
return snpy_generator
def write_ped(self, dirfilespheno, outputfile, pheno):
"""
read the file inside a folder and parse to write a .ped.
"""
for i in os.listdir(dirfilespheno):
all_rs_ind_tmp = np.array(
["0 0"] * len(self.all_rs),
dtype='S3') #initialaze every genotype with "0 0"
sex = ""
all_names = []
all_gen = []
print dirfilespheno + "/" + i,
if "XY" in i:
sex = "1"
elif "XX" in i:
sex = "2"
else:
sex = "9" #sex="other"
try:
snps = sn.parse(dirfilespheno + "/" + i) #take another files
except:
print " ERRO 1"
continue
try:
for cur_snp in snps:
if len(
cur_snp.genotype
) == 2 and cur_snp.genotype in self.valid_alleles: # "--" and cur_snp.genotype != "II" and cur_snp.genotype != "DI":
all_names.append(cur_snp.name)
all_gen.append(
"%s %s" %
(cur_snp.genotype[0], cur_snp.genotype[1]))
except:
print " ERRO 2"
continue
try:
idx = np.flatnonzero(np.in1d(self.all_rs, np.array(all_names)))
except:
print " ERRO 3"
continue
print ""
all_rs_ind_tmp[idx] = np.array(all_gen)
outputfile.write(
str(self.Family_ID) + " " + str(self.Individual_ID) + " " +
"0 0 " + sex + " " + pheno + " ")
for i in all_rs_ind_tmp:
outputfile.write(i + " ")
outputfile.write("\n")
self.Family_ID = self.Family_ID + 1
self.Individual_ID = self.Individual_ID + 1
def convert_genotype_to_hdf5(csv_content, output_file, source=None):
log.info('Convert genotype from text format to HDF5 format %s' % (output_file))
# guess the source
fd = io.StringIO(csv_content)
if source is None:
source = sn.guess_source_from_content(csv_content)
version = ''
snps = sn.parse(fd, source)
start_chr = None
f = h5py.File(output_file, 'w')
pos_snps = []
num_snps = 0
group = None # Eliminates syntax error highlight in Eclipse.
for snp in snps:
num_snps += 1
if snp.chromosome != start_chr:
if start_chr is not None:
sorted(pos_snps, key=lambda x: x[0])
positions, ids, snps = zip(*pos_snps)
group.create_dataset('ids', (len(positions),), chunks=True, compression='lzf', dtype='S15', data=ids)
group.create_dataset('positions', (len(positions),), chunks=True, compression='lzf', dtype='i8', data=positions)
group.create_dataset('snps', (len(positions),), chunks=True, compression='lzf', dtype='S2', data=snps)
pos_snps = []
start_chr = snp.chromosome
group = f.create_group("Chr%s" % start_chr)
pos_snps.append((int(snp.position), snp.name.encode('utf-8'), snp.genotype.encode('utf-8')))
sorted(pos_snps, key=lambda x: x[0])
positions, ids, snps = zip(*pos_snps)
group.create_dataset('ids', (len(positions),), chunks=True, compression='lzf', dtype='S15', data=ids)
group.create_dataset('positions', (len(positions),), chunks=True, compression='lzf', dtype='i8', data=positions)
group.create_dataset('snps', (len(positions),), chunks=True, compression='lzf', dtype='S2', data=snps)
# find out the version
f.attrs['source'] = source
# based on https://en.wikipedia.org/wiki/23andMe
if source == '23andme':
if num_snps <= 576000:
version = 'v1'
elif num_snps <= 597000:
version = 'v2'
# TODO current workaround because oauth genotype are > 1000000
elif num_snps <= 611000 or num_snps > 1000000:
version = 'v4'
elif num_snps <= 992000:
version = 'v3'
f.attrs['version'] = version
f.attrs['gender'] = 'm' if 'ChrY' in f.keys() else 'f'
f.close()
def preprocess_genotype_tsv(input, source, output):
"""
Preprocess input files from other source.
:param input:
:param source:
:param output:
:return:
"""
snps = sn.parse(input, source)
with open(output, 'w') as fh:
for s in snps:
rs = s.name
if rs[0:2] != 'rs':
continue
ref, alt = split_genotype(s.genotype)
if alt is None:
continue
fh.write('\t'.join([rs, ref, alt]) + '\n')
return output
def update_index(seq_path, seq_id, dbsnp=None, index_root=None):
"""Update 'consequence' index with SNPs from a given file.
Currently a file should either be in VCF, with a filename like
``*.vcf``, or in one of openSNP-provided formats, in that case,
the name is expected to be left unmodified:
``user_id.format.submission_id``.
If `dbsnp` argument is provided, only SNPs contained in `dbsnp`
will be indexed.
"""
g = Genome(base_path=index_root)
dbsnp = {} if not dbsnp else \
dict((r.POS, r.ID) for r in vcf.VCFReader(open(dbsnp, "rb")))
for record in sn.parse(seq_path,
source="vcf" if seq_path.endswith(".vcf") else None):
# XXX here comes the punchline, VCF uses 1-based indexing,
# so does SAM / BAM, but 'pysam' converts *all* indexes to
# 0-based!
chrom, pos = key = record.chromosome, record.position - 1
if dbsnp and pos not in dbsnp:
continue
elif key not in g:
# Note(Sergei): dbnSNP name defaults to '*'.
g[key] = Chunk(dbsnp.get(pos, record.name or "*"))
for alt in record.genotype:
known = getattr(g[key], alt, None) or set()
if seq_id in known:
continue
known.add(seq_id)
g[key] = g[key]._replace(**{alt: known})
g.dump()
def creat_map(output_file="output.map"):
"""Create the .map file"""
input_file=raw_input("\n-------folders with files: ")
dir_list=input_file.split(";")
files=[]
file_dir={}
for i in dir_list:
files.extend(os.listdir(i)) #all files names in a dir
for j in os.listdir(i):
file_dir[j]=i
#print j,i
rs_list=[]
map_list=[]
idx=0
print "\nParse the file ", file_dir[files[0]]+"/"+files[0],
try:
snps = sn.parse(file_dir[files[0]]+"/"+files[0])#take the first file
for i in snps: #initialaze rs_list and map_list with the first file
rs_list.append(i[0])
map_list.append((i[0],i[2],i[3]))
except:
print " ERRO 1"
print " OK"
dtype = [('rs', 'S10'), ('chrom', 'S10'), ('position', int)]
map_list = np.array( map_list, dtype=dtype)
for j in files[1:]:
print "Parse the file ",file_dir[j]+"/"+j,
try:
snps = sn.parse(file_dir[j]+"/"+j) #take another files
rs_list_tmp=[]
map_list_tmp=[]
except:
print " ERRO 2"
continue
try:
for i in snps:
rs_list_tmp.append(i[0])
map_list_tmp.append((i[0],i[2],i[3]))
except:
print " ERRO 3"
continue
try: #erro 3 in this files for exemplo
rs_list_dif=np.setdiff1d( rs_list_tmp, rs_list) #user36_file207_yearofbirth_1986_sex_XY.23andme-exome-vcf.txt
except: #user36_file327_yearofbirth_1986_sex_XY.23andme-exome-vcf.txt
print " ERRO 4"
continue
idx = np.flatnonzero(np.in1d(np.array(rs_list_tmp),np.array(rs_list_dif))) #return the index of the elements of "all_names" in "all_rs"
map_list_tmp=np.array(map_list_tmp, dtype=dtype)
map_list=np.array(np.concatenate(( map_list, map_list_tmp[idx] )) , dtype=dtype)
print " OK"
print "sort..."
map_list = np.sort(map_list, order=['chrom', 'position'])
chrommosomos=["1","2","3","4","5","6","7","8","9","10","11","12","13","14","15","16","17","18","19","20","21","22","X","Y","MT"]
ofile = open(output_file,'w') # open file for writing
print "write the output file..."
for i in chrommosomos:
if i in map_list['chrom']:
idx = map_list['chrom']==i
for i in map_list[:][idx]:
ofile.write(str(i[1])+" "+str(i[0])+" "+str(0)+" "+str(i[2])+"\n")
ofile.close()
sex=""
print i
if "XY" in i:
sex="1"
print "m"
elif "XX" in i:
sex="2"
print "f"
else:
sex="other"
print "o"
snps = sn.parse(dir_files_phenotype1+"/"+i) #take another files
all_names=[]
all_gen=[]
for cur_snp in snps:
if len(cur_snp.genotype)==2 and cur_snp.genotype != "--":
all_names.append(cur_snp.name)
all_gen.append("%s %s" % (cur_snp.genotype[0], cur_snp.genotype[1]))
idx = np.flatnonzero(np.in1d(all_rs, np.array(all_names))) #return the index of the elements of "all_names" in "all_rs"
all_rs_ind_tmp[idx] = np.array(all_gen)
def create_map(self, folders_case, folders_control):
"""Create the .map file"""
print "\n\nMAP FILE (", time.asctime(), ")\n"
dir_list = folders_case[:]
dir_list.extend(folders_control)
files = []
file_dir = {}
map_list = []
idx = 0
for i in dir_list:
files.extend(os.listdir(i)) #it get all files names in a dir
for j in os.listdir(i):
file_dir[j] = i #dictionari with file:dir
print "Reading the files:\n\n", file_dir[
files[0]] + "/" + files[0], #parse the first file
try:
snps = sn.parse(file_dir[files[0]] + "/" +
files[0]) #take the first file
for i in snps: #initialaze rs_list and map_list with the first file
map_list.append((i[0], i[2], i[3]))
except:
print " ERRO 1"
print ""
dtype = [('rs', 'S10'), ('chrom', 'S10'), ('position', int)]
map_list = np.array(map_list, dtype=dtype)
for j in files[1:]:
map_list_tmp = []
print file_dir[j] + "/" + j,
try:
snps = sn.parse(file_dir[j] + "/" + j) #take another files
except:
print " ERRO 2"
continue
try:
for i in snps:
map_list_tmp.append((i[0], i[2], i[3]))
except:
print " ERRO 3"
continue
print ""
map_list_tmp = np.array(map_list_tmp, dtype=dtype)
map_list = np.array(np.concatenate((map_list, map_list_tmp)),
dtype=dtype)
u, indices = np.unique(map_list['rs'], return_index=True)
map_list = map_list[indices]
array_chrom = np.unique(
map_list['chrom']
) #add new elements to end of the self.chrommosomos
idx_chr = np.in1d(array_chrom, self.chrommosomos)
self.chrommosomos = np.concatenate(
(self.chrommosomos, array_chrom[idx_chr == False]))
map_list = np.sort(map_list, order=['chrom', 'position'])
ofile = open(self.outputmap, 'w') # open file for writing
print "there are", len(
map_list['rs']), "SNPs.\nwriting the", self.outputmap, "file..."
for i in self.chrommosomos:
if i in map_list['chrom']:
idx = map_list['chrom'] == i
for i in map_list[:][idx]:
ofile.write(
str(i[1]) + " " + str(i[0]) + " " + str(0) + " " +
str(i[2]) + "\n")
ofile.close()
