def main():
parser = argparse.ArgumentParser(
description="Extract the SFS from a binch of merged SLiM output files")
parser.add_argument(
"-i",
"--input",
required=True,
dest="input",
type=str,
help="The name of the input file (or the input directory)")
parser.add_argument(
"-o",
"--output",
required=True,
dest="output",
type=str,
help="The name of the output file you want to write to")
parser.add_argument(
"-d",
"--dir",
required=False,
dest="dir",
action='store_true',
help=
"Use this flag if you want to combine a number of files in the same directory",
default=False)
args = parser.parse_args()
if args.dir:
files = glob.glob(args.input + '/*sfs')
else:
files = [args.input]
sites = float(len(files)) * 1e6
# Get a list of dictionaries containing the SFSs
dicts = [getDict(i) for i in files]
nonsynSFS, synSFS = mergeDicts(dicts)
nonsynSFS[0] += (sites * 0.75) - sum(nonsynSFS[1:])
synSFS[0] += (sites * 0.25) - sum(synSFS[1:])
ds = float(synSFS[-1]) / sum(synSFS)
dn = float(nonsynSFS[-1]) / sum(nonsynSFS)
print 'nonsyn pi:', pi(nonsynSFS)
print 'syn pi:', pi(synSFS)
print 'pi / pi_0:', pi(synSFS) / 0.01
print 'dN', dn
print 'ds', ds
print 'adv. Cont', float(sum(mergeTheSFS([
i['m2'] for i in dicts
])[1:-1])) / (sum(mergeTheSFS([i['m1'] for i in dicts])[1:-1]) +
sum(mergeTheSFS([i['m2'] for i in dicts])[1:-1]))
print 'dN/dS:', dn / ds
def main():
parser = argparse.ArgumentParser(
description="Extract the SFS from a binch of merged SLiM output files")
parser.add_argument(
"-i",
"--input",
required=True,
dest="input",
type=str,
help="The name of the directory containing the SLiM output")
parser.add_argument(
"-o",
"--output",
required=True,
dest="output",
type=str,
help="The name of the output file you want to write to")
args = parser.parse_args()
# output = []
for m in range(1, 9):
print 'm' + str(m)
full_sfs = []
for i in glob.glob(args.input + '/R*'):
num = i.split('/')[-1].split('.')[2]
process = subprocess.Popen(['zgrep', 'm' + str(m), i],
stdout=subprocess.PIPE).communicate()[0]
fixations, sfs = getSFSfromSLiM(process)
if fixations == None: continue
polymorphs = sum(sfs)
sfs[-1] += fixations
sfs[0] = 0
print SFS.pi(sfs) / 140000.
print ':'.join(map(str, sfs))
continue
if len(full_sfs) == 0:
full_sfs = sfs
else:
full_sfs = SFS.merge_SFS(full_sfs, sfs)
# output.append(['m'+str(m),full_sfs])
return
txt = open(args.output, 'w')
for i in output:
print i
txt.write(i[0] + '\n')
txt.write(' '.join(map(str, i[1])) + '\n')
txt.close()
def main():
parser = argparse.ArgumentParser(
description=
"Combine all the sfs files coming out of the sfs_from_slim_update_bootstrap.py script"
)
parser.add_argument(
"-i",
"--input",
required=True,
dest="input",
type=str,
help="The name of the file that contains the sfs files")
parser.add_argument("-o",
"--output",
required=True,
dest="output",
type=str,
help="The name of the output file")
args = parser.parse_args()
sfs_dict = {}
for i in gzip.open(args.input):
z = i.split('[')
region = z[1].strip("'").replace("'", '')
sfs_temp = map(
int, z[2].replace(']', '').replace(',', '').strip().split(' '))
try:
sfs_dict[region].append(sfs_temp)
except KeyError:
sfs_dict[region] = [sfs_temp]
data = []
for i in sfs_dict.keys():
sfs = sfs_dict[i][0]
for j in sfs_dict[i][1:]:
sfs = SFS_tools.merge_SFS(sfs, j)
stream = i.split('.')[0]
dist = map(int, i.replace(',', '').split('.')[1].split('-'))
if stream == 'u':
mult = -1
else:
mult = 1
mid = mult * sum(dist) / 2
data.append([mid, SFS_tools.pi(sfs), SFS_tools.tajima(sfs)])
pd.DataFrame(data, columns=['dist', 'pi', 'TD']).to_csv(args.output)
def main():
parser = argparse.ArgumentParser(
description=
"Give a directory and I'll analyse the patterns of diversity around a simulated exon"
)
parser.add_argument(
"-i",
"--input",
required=True,
dest="input",
type=str,
help="The name of the file that contains the SLiM output")
args = parser.parse_args()
count = 0
r_bins = [float(i) / 10
for i in range(1, 500, 1)] + [1. * i for i in range(50, 500)]
approx = combinedSel(np.array(r_bins))
pylab.plot(r_bins, approx * 0.01, 'r')
for i in glob.glob(args.input + '*.out.gz'):
if count == 0:
data = processSLiM(i)
else:
temp = processSLiM(i)
data = pd.concat([data, temp])
count += 1
if count == 50:
break
top = max(range(len(r_bins)))
analysis = []
for r in range(len(r_bins)):
bin = r_bins[r]
if r == top:
next = 1e8
else:
next = r_bins[r + 1]
in_range = list(data[(data.r_dist_true >= bin)
& (data.r_dist_true < next)].sfs)
if len(in_range) == 0: continue
analysis.append([r_bins[r], SFS.pi(mergeManySFS(in_range))])
true_r = pd.DataFrame(analysis, columns=['dist', 'pi'])
pylab.plot(true_r.dist, true_r.pi, 'b')
pylab.show()
def get_summary(sfs_dict, label):
out = []
for key in sfs_dict.keys():
mid = key
if sum(sfs_dict[key][1:-1]) == 0:
out.append([
label, mid, 'Nan', 'Nan', 'Nan', 'Nan', 'Nan', 'Nan',
sum(sfs_dict[key])
])
elif sum(sfs_dict[key][1:-1]) > 0:
out.append([
label,
mid,
SFS.pi(sfs_dict[key]),
SFS.xsi(sfs_dict[key]),
SFS.pi2(sfs_dict[key]),
SFS.fwh(sfs_dict[key]),
SFS.theta_W(sfs_dict[key]),
SFS.tajima(sfs_dict[key]),
# SFS.KZl(sfs_dict[key]),
sum(sfs_dict[key])
])
return out
def main():
parser = argparse.ArgumentParser(
description="Takes the summary file of recombination distances and ")
parser.add_argument("-i",
"--input",
required=True,
dest="input",
type=str,
help="The name of the sorted, bed file of segments")
parser.add_argument("-o",
"--output",
required=True,
dest="output",
type=str,
help="The name of the output file")
parser.add_argument(
"-l",
"--label",
required=False,
dest="label",
type=str,
help=
"Add a label to this data, e;g; which chromosome do they come from? The default will be 'Autosomes'",
default='Autosomes')
parser.add_argument(
"--ncpg",
required=False,
dest="ncpg",
action='store_true',
help="Add this flag if you want to analyse the non-CpG sites",
default=False)
parser.add_argument("--cne",
required=False,
dest="cne",
action='store_true',
help="Add this flag if you are analysing CNEs",
default=False)
parser.add_argument("--Cox",
required=False,
dest="Cox",
action='store_true',
help="Add this flag if you want to use the Cox map",
default=False)
parser.add_argument(
"--GC",
required=False,
dest="GC",
action='store_true',
help=
"Add this flag if you want to include GeneConversion according to the Paigen et al estimates",
default=False)
args = parser.parse_args()
data = pd.read_csv(args.input, compression='gzip', header=None, sep='\t')
data['scale'] = [-1 if x.split('.')[0] == 'u' else 1 for x in data[0]]
# The following gene conversion parameters come from Paigen et al 2008 PLoS Genetics
nc_gc_ratio = 0.105 # The relative rate of NCO gene conversion compared to CO Gene Conversion 0.105 for Paigen
tract_length = 144 # average tract length
if args.Cox:
data['r_co'] = data[7] * 426200 * 4
# RecPos is the index of the dataframe where recombination rates are held
recPos = 'r_co'
data['r_gc'] = 426200 * 4 * (nc_gc_ratio * data[7] /
data[3]) * tract_length * (1 - np.exp(
(-1. * data[3]) / tract_length))
data['joint'] = data['r_gc'] + data['r_co']
if args.GC:
recPos = 'joint'
else:
data['r_co'] = data[2]
recPos = 'r_co'
data['r_gc'] = (nc_gc_ratio * data[2] /
data[3]) * tract_length * (1 - np.exp(
(-1. * data[3]) / tract_length))
data['joint'] = data['r_gc'] + data['r_co']
if args.GC:
recPos = 'joint'
data['dist'] = data[recPos] * data['scale']
if args.cne:
roundBy = 2
bins = np.logspace(0, 2.477122, 100) - 1
# bins = range(0, 200, roundBy) ## For CNEs
else:
roundBy = 30
bins = np.logspace(0, 3.477122, 100) - 1
# bins = range(0, 3000, roundBy) ## For Exons
output_lines = []
output_lines_2 = []
for i in range(len(bins)):
if i < 99:
chunk = data.loc[(data[recPos] >= bins[i])
& (data[recPos] < bins[i + 1])]
else:
chunk = data.loc[(data[recPos] >= bins[i])]
if len(chunk) == 0: continue
SFS, div = summariseChunk(chunk, ncpg=args.ncpg)
up_chunk = chunk[chunk['dist'] < 0]
if len(up_chunk) == 0: continue
SFS_up, div_up = summariseChunk(up_chunk, ncpg=args.ncpg)
down_chunk = chunk[chunk['dist'] > 0]
if len(down_chunk) == 0: continue
SFS_down, div_down = summariseChunk(down_chunk, ncpg=args.ncpg)
if sum(SFS) == 0: continue
else:
outline = [
bins[i],
args.label,
SFS_tools.pi(SFS),
pgm.jukes_cantor(float(div[0]) / sum(SFS)),
pgm.jukes_cantor(float(div[1]) / sum(SFS)),
SFS_tools.tajima(SFS),
sum(SFS),
]
if sum(SFS_up) == 0: continue
else:
outline_up = [
-1 * bins[i],
args.label,
SFS_tools.pi(SFS_up),
pgm.jukes_cantor(float(div_up[0]) / sum(SFS_up)),
pgm.jukes_cantor(float(div_up[1]) / sum(SFS_up)),
SFS_tools.tajima(SFS_up),
sum(SFS_up),
]
if sum(SFS_down) == 0: continue
else:
outline_down = [
bins[i],
args.label,
SFS_tools.pi(SFS_down),
pgm.jukes_cantor(float(div_down[0]) / sum(SFS_down)),
pgm.jukes_cantor(float(div_down[1]) / sum(SFS_down)),
SFS_tools.tajima(SFS_down),
sum(SFS_down),
]
output_lines.append(outline)
output_lines_2.append(outline_up)
output_lines_2.append(outline_down)
output1 = pd.DataFrame(output_lines,
columns=[
'mid', 'label', 'pi', 'fam_div_jc',
'rat_div_jc', 'tajima', 'sites'
])
output1.to_csv(args.output)
output2 = pd.DataFrame(output_lines_2,
columns=[
'mid', 'label', 'pi', 'fam_div_jc',
'rat_div_jc', 'tajima', 'sites'
])
output2.to_csv('split_' + args.output)
for i in glob.glob('*/*sfs'):
x = open(i).readlines()
name = i.split('/')[0]
if name[0] == 'u':
mult = -1
elif name[0] == 'd':
mult = 1
interval = name.split('.')[1]
dist = mult * (int(interval.split('-')[0]) +
int(interval.split('-')[1])) / 2
x = open(i).readlines()
sfs = map(float, x[0].strip().split(','))
if sfs == [0.0]:
continue
outline = [
dist,
sum(sfs),
SFS.pi(sfs),
SFS.xsi(sfs),
SFS.pi2(sfs),
SFS.theta_W(sfs),
SFS.fwh(sfs),
SFS.tajima(sfs), sfs[-1] / sum(sfs)
]
out_lines.append(outline)
# output.write(','.join(map(str,outline))+'\n')
data = pd.DataFrame(out_lines, columns=headings)
data.sort_values('dist').to_csv(sys.argv[1])
def main():
parser = argparse.ArgumentParser(
description="Takes the summary file of recombination distances and ")
parser.add_argument("-i",
"--input",
required=True,
dest="input",
type=str,
help="The name of the sorted, bed file of segments")
parser.add_argument("-o",
"--output",
required=True,
dest="output",
type=str,
help="The name of the output file")
parser.add_argument(
"-l",
"--label",
required=False,
dest="label",
type=str,
help=
"Add a label to this data, e;g; which chromosome do they come from? The default will be 'Autosomes'",
default='Autosomes')
parser.add_argument(
"--ncpg",
required=False,
dest="ncpg",
action='store_true',
help="Add this flag if you want to analyse the non-CpG sites",
default=False)
parser.add_argument(
"--dir",
required=False,
dest="dir",
action='store_true',
help="Add this flag if you are analysing a directory of files",
default=False)
args = parser.parse_args()
if not args.dir:
data = pd.read_csv(args.input, header=None, sep='\t')
else:
data = pd.concat([
pd.read_csv(i, header=None, sep='\t')
for i in glob.glob(args.input + '*')
])
ranges = range(0, 100, 1) ## For CNEs
# ranges = range(0, 3000, 30) ## For Exons
output_lines = []
for i in range(len(ranges)):
if i < 99:
chunk = data.loc[(data[2] >= ranges[i])
& (data[2] < ranges[i + 1])]
else:
chunk = data.loc[(data[2] >= ranges[i])]
if len(chunk) == 0: continue
SFS, div = summariseChunk(chunk, ncpg=args.ncpg)
if sum(SFS) == 0: continue
outline = [
ranges[i],
args.label,
SFS_tools.pi(SFS),
pgm.jukes_cantor(float(div[0]) / sum(SFS)),
pgm.jukes_cantor(float(div[1]) / sum(SFS)),
SFS_tools.tajima(SFS),
sum(SFS),
]
output_lines.append(outline)
output = pd.DataFrame(output_lines,
columns=[
'distance', 'label', 'pi', 'fam_div_jc',
'rat_div_jc', 'tajima', 'sites'
])
output.to_csv(args.output)
def summariseChunk(chunk): recDist = chunk[1].mean() sfs = [] for i in list(chunk.columns)[2:]: sfs.append( chunk[i].sum() ) return [recDist, SFS_tools.pi(sfs)]