def plot_snp_dens_distrib(variations, window_size, data_dir, write_bg=False):
# Calculate and plot variations density distribution
density = calc_snp_density(variations, window_size)
density_distrib, bins = histogram(density, 20)
fpath = join(data_dir, 'snps_density.png')
title = 'SNP density distribution per {} bp windows'.format(window_size)
plot_distrib(density_distrib, bins, fhand=open(fpath, 'w'), color='c',
mpl_params={'set_xlabel': {'args': ['SNP density'],
'kwargs': {}},
'set_ylabel': {'args': ['SNP number'],
'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}},
'set_yscale': {'args': ['log'], 'kwargs': {}}})
# Manhattan plot for SNP density
fpath = join(data_dir, 'snps_density_manhattan.png')
fhand = open(fpath, 'w')
title = 'SNP denisity along the genome'
chrom = _load_matrix(variations, CHROM_FIELD)
pos = _load_matrix(variations, POS_FIELD)
manhattan_plot(chrom, pos, density,
mpl_params={'set_xlabel': {'args': ['Chromosome'],
'kwargs': {}},
'set_ylabel': {'args': ['SNP per {} bp'.format(window_size)],
'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}}},
fhand=fhand, figsize=(15, 7.5), ylim=1)
# Save in bedgraph format
if write_bg:
bg_fhand = open(join(data_dir, 'snp_density.bg'), 'w')
pos_dens = PositionalStatsCalculator(chrom, pos, density)
pos_dens.write(bg_fhand, 'snp_density',
'SNP number in {} bp around'.format(window_size),
track_type='bedgraph')
def plot_maf_depth(variations, data_dir, min_depth=DEF_MIN_DEPTH,
chunk_size=SNPS_PER_CHUNK):
maf_dp_distribs = calc_maf_depth_distribs_per_sample(variations,
min_depth=min_depth,
n_bins=100,
chunk_size=SNPS_PER_CHUNK)
maf_dp_distribs, bins = maf_dp_distribs
maf_dp_dir = os.path.join(data_dir, 'maf_depth')
if not os.path.exists(maf_dp_dir):
os.mkdir(maf_dp_dir)
samples = variations.samples
if samples is None:
samples = range(maf_dp_distribs.shape[0])
for sample, distrib in zip(samples, maf_dp_distribs):
fpath = join(maf_dp_dir, '{}.png'.format(sample))
title = 'Depth based Maximum allele frequency (MAF) distribution {}'
title = title.format(sample)
mpl_params = {'set_xlabel': {'args': ['MAF (depth)'], 'kwargs': {}},
'set_ylabel': {'args': ['SNPs number'], 'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}},
'set_yscale': {'args': ['log'], 'kwargs': {}}}
plot_distrib(distrib, bins, fhand=open(fpath, 'w'), figsize=(10, 10),
mpl_params=mpl_params, n_ticks=10)
def plot_hwe(variations, max_num_alleles, data_dir, ploidy=2,
min_num_genotypes=MIN_NUM_GENOTYPES_FOR_POP_STAT,
chunk_size=SNPS_PER_CHUNK):
fpath = join(data_dir, 'hwe_chi2_distrib.png')
fhand = open(fpath, 'w')
fig = Figure(figsize=(10, 20))
canvas = FigureCanvas(fig)
num_alleles = range(2, max_num_alleles + 1)
gs = gridspec.GridSpec(len(num_alleles), 1)
for i, num_allele in enumerate(num_alleles):
df = len(list(combinations_with_replacement(range(num_allele),
ploidy))) - num_allele
hwe_test = calc_hwe_chi2_test(variations, num_allele=num_allele,
min_num_genotypes=min_num_genotypes,
chunk_size=chunk_size)
hwe_chi2 = hwe_test[:, 0]
hwe_chi2_distrib, bins = histogram(hwe_chi2, n_bins=50)
# Plot observed distribution
axes = fig.add_subplot(gs[i, 0])
title = 'Chi2 df={} statistic values distribution'.format(df)
mpl_params = {'set_xlabel': {'args': ['Chi2 statistic'], 'kwargs': {}},
'set_ylabel': {'args': ['SNP number'], 'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}}}
plot_distrib(hwe_chi2_distrib, bins, axes=axes, mpl_params=mpl_params)
# Plot expected chi2 distribution
axes = axes.twinx()
rv = chi2(df)
x = numpy.linspace(0, max(hwe_chi2), 1000)
axes.plot(x, rv.pdf(x), color='b', lw=2, label='Expected Chi2')
axes.set_ylabel('Expected Chi2 density')
canvas.print_figure(fhand)
def plot_missing_gt_rate_per_snp(variations,
data_dir,
chunk_size=SNPS_PER_CHUNK):
_calc_missing_gt = partial(calc_missing_gt, rates=True, axis=1)
distrib, bins = histogram_for_chunks(variations,
calc_funct=_calc_missing_gt,
range_=(0, 1),
n_bins=20,
chunk_size=chunk_size)
fpath = join(data_dir, 'missing_gt_rate.png')
title = 'Missing Genotype rates per SNP distribution'
plot_distrib(distrib,
bins,
fhand=open(fpath, 'w'),
color='c',
mpl_params={
'set_xlabel': {
'args': ['Missing GT rate'],
'kwargs': {}
},
'set_ylabel': {
'args': ['SNP number'],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
}
})
def plot_call_field_distribs_per_gt_type(variations, field, max_value,
data_dir, chunk_size=SNPS_PER_CHUNK):
# Field distribution per sample
field_name = field.split('/')[-1]
fpath = join(data_dir, '{}_distribution_per_sample.png'.format(field_name))
mask_funcs = [call_is_het, call_is_hom]
names = ['Heterozygous', 'Homozygous']
distribs = []
for mask_func in mask_funcs:
dp_distribs, bins = calc_field_distribs_per_sample(variations,
field=field,
range_=(0, max_value),
n_bins=max_value,
chunk_size=chunk_size,
mask_func=mask_func,
mask_field=GT_FIELD)
distribs.append(dp_distribs)
title = '{} distribution per sample'.format(field_name)
mpl_params = {'set_xlabel': {'args': ['Samples'], 'kwargs': {}},
'set_ylabel': {'args': [field_name], 'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}}}
figsize = (variations[GT_FIELD].shape[1], 7)
plot_boxplot_from_distribs_series(distribs, fhand=open(fpath, 'w'),
mpl_params=mpl_params, figsize=figsize,
colors=['pink', 'tan'],
labels=names,
xticklabels=variations.samples)
# Overall field distributions
fpath = join(data_dir, '{}_distribution.png'.format(field_name))
fhand = open(fpath, 'w')
fig = Figure(figsize=(20, 15))
canvas = FigureCanvas(fig)
i = 1
for distrib, name in zip(distribs, names):
distrib = numpy.sum(dp_distribs, axis=0)
distrib_cum = calc_cum_distrib(distrib)
axes = fig.add_subplot(len(names) * 100 + 20 + i)
i += 1
title = '{} distribution all samples {}'.format(field_name, name)
plot_distrib(distrib, bins, axes=axes,
mpl_params={'set_xlabel': {'args': [field_name],
'kwargs': {}},
'set_ylabel': {'args': ['Number of GTs'],
'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}}})
distrib_cum = distrib_cum/distrib_cum[0] * 100
axes = fig.add_subplot(len(names) * 100 + 20 + i)
i += 1
title = '{} cumulative distribution all samples {}'.format(field_name,
name)
plot_distrib(distrib_cum, bins, axes=axes,
mpl_params={'set_xlabel': {'args': [field_name],
'kwargs': {}},
'set_ylabel': {'args': ['% calls > Depth '],
'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}}})
canvas.print_figure(fhand)
def plot_hwe(variations,
max_num_alleles,
data_dir,
ploidy=2,
min_num_genotypes=MIN_NUM_GENOTYPES_FOR_POP_STAT,
chunk_size=SNPS_PER_CHUNK):
fpath = join(data_dir, 'hwe_chi2_distrib.png')
fhand = open(fpath, 'w')
fig = Figure(figsize=(10, 20))
canvas = FigureCanvas(fig)
num_alleles = range(2, max_num_alleles + 1)
gs = gridspec.GridSpec(len(num_alleles), 1)
for i, num_allele in enumerate(num_alleles):
df = len(list(combinations_with_replacement(range(num_allele),
ploidy))) - num_allele
hwe_test = calc_hwe_chi2_test(variations,
num_allele=num_allele,
min_num_genotypes=min_num_genotypes,
chunk_size=chunk_size)
hwe_chi2 = hwe_test[:, 0]
hwe_chi2_distrib, bins = histogram(hwe_chi2, n_bins=50)
# Plot observed distribution
axes = fig.add_subplot(gs[i, 0])
title = 'Chi2 df={} statistic values distribution'.format(df)
mpl_params = {
'set_xlabel': {
'args': ['Chi2 statistic'],
'kwargs': {}
},
'set_ylabel': {
'args': ['SNP number'],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
}
}
plot_distrib(hwe_chi2_distrib, bins, axes=axes, mpl_params=mpl_params)
# Plot expected chi2 distribution
axes = axes.twinx()
rv = chi2(df)
x = numpy.linspace(0, max(hwe_chi2), 1000)
axes.plot(x, rv.pdf(x), color='b', lw=2, label='Expected Chi2')
axes.set_ylabel('Expected Chi2 density')
canvas.print_figure(fhand)
def plot_missing_gt_rate_per_snp(variations, data_dir,
chunk_size=SNPS_PER_CHUNK):
_calc_missing_gt = partial(calc_missing_gt, rates=True, axis=1)
distrib, bins = histogram_for_chunks(variations,
calc_funct=_calc_missing_gt,
range_=(0, 1), n_bins=20,
chunk_size=chunk_size)
fpath = join(data_dir, 'missing_gt_rate.png')
title = 'Missing Genotype rates per SNP distribution'
plot_distrib(distrib, bins, fhand=open(fpath, 'w'), color='c',
mpl_params={'set_xlabel': {'args': ['Missing GT rate'],
'kwargs': {}},
'set_ylabel': {'args': ['SNP number'],
'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}}})
def plot_maf_depth(variations,
data_dir,
min_depth=DEF_MIN_DEPTH,
chunk_size=SNPS_PER_CHUNK):
maf_dp_distribs = calc_maf_depth_distribs_per_sample(
variations, min_depth=min_depth, n_bins=100, chunk_size=SNPS_PER_CHUNK)
maf_dp_distribs, bins = maf_dp_distribs
maf_dp_dir = os.path.join(data_dir, 'maf_depth')
if not os.path.exists(maf_dp_dir):
os.mkdir(maf_dp_dir)
samples = variations.samples
if samples is None:
samples = range(maf_dp_distribs.shape[0])
for sample, distrib in zip(samples, maf_dp_distribs):
fpath = join(maf_dp_dir, '{}.png'.format(sample))
title = 'Depth based Maximum allele frequency (MAF) distribution {}'
title = title.format(sample)
mpl_params = {
'set_xlabel': {
'args': ['MAF (depth)'],
'kwargs': {}
},
'set_ylabel': {
'args': ['SNPs number'],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
},
'set_yscale': {
'args': ['log'],
'kwargs': {}
}
}
plot_distrib(distrib,
bins,
fhand=open(fpath, 'w'),
figsize=(10, 10),
mpl_params=mpl_params,
n_ticks=10)
def plot_inbreeding_coefficient(variations, max_num_allele, data_dir,
window_size, chunk_size=SNPS_PER_CHUNK,
min_num_genotypes=MIN_NUM_GENOTYPES_FOR_POP_STAT,
write_bg=False, calc_genome_wise=False):
# Calculate Inbreeding coefficient distribution
inbreed_coef = calc_inbreeding_coef(variations, chunk_size=chunk_size,
min_num_genotypes=min_num_genotypes)
ic_distrib, bins = histogram(inbreed_coef, 50, range_=(-1, 1))
fpath = join(data_dir, 'inbreeding_coef_distribution.png')
fhand = open(fpath, 'w')
title = 'Inbreeding coefficient distribution all samples'
plot_distrib(ic_distrib, bins, fhand=fhand,
mpl_params={'set_xlabel': {'args': ['Inbreeding coefficient'],
'kwargs': {}},
'set_ylabel': {'args': ['Number of SNPs'],
'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}},
'set_xlim': {'args': [-1, 1], 'kwargs': {}}})
# Save in bedgraph file
if calc_genome_wise:
bg_fhand = open(join(data_dir, 'ic.bg'), 'w')
chrom = _load_matrix(variations, CHROM_FIELD)
pos = _load_matrix(variations, POS_FIELD)
pos_ic = PositionalStatsCalculator(chrom, pos, inbreed_coef)
if write_bg:
pos_ic.write(bg_fhand, 'IC', 'Inbreeding coefficient',
track_type='bedgraph')
# Plot Ic along genome taking sliding windows
pos_ic = pos_ic.calc_window_stat()
chrom, pos, ic_windows = pos_ic.chrom, pos_ic.pos, pos_ic.stat
fpath = join(data_dir, 'ic_manhattan.png')
fhand = open(fpath, 'w')
title = 'Inbreeding coefficient (IC) along the genome'
manhattan_plot(chrom, pos, ic_windows, fhand=fhand, figsize=(15, 7.5),
ylim=-1,
mpl_params={'set_xlabel': {'args': ['Chromosome'],
'kwargs': {}},
'set_ylabel': {'args': ['IC'],
'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}}})
def plot_obs_het(variations, data_dir, chunk_size=SNPS_PER_CHUNK,
min_num_genotypes=MIN_NUM_GENOTYPES_FOR_POP_STAT):
# Calculate observed heterozygosity distribution by snp
_calc_obs_het_by_var = partial(calc_obs_het,
min_num_genotypes=min_num_genotypes)
distrib = histogram_for_chunks(variations, calc_funct=_calc_obs_het_by_var,
n_bins=25, range_=(0, 1),
chunk_size=chunk_size)
obs_het_var_distrib, bins1 = distrib
# Calculate observed heterozygosity distribution by sample
obs_het_by_sample = calc_obs_het_by_sample(variations,
chunk_size=chunk_size)
obs_het_sample_distrib, bins2 = histogram(obs_het_by_sample, n_bins=25,
range_=(0, 1))
# Plot distributions
fpath = join(data_dir, 'obs_het.png')
fhand = open(fpath, 'w')
fig = Figure(figsize=(10, 10))
canvas = FigureCanvas(fig)
axes = fig.add_subplot(211)
title = 'SNP observed Heterozygosity distribution'
plot_distrib(obs_het_var_distrib, bins=bins1, fhand=open(fpath, 'w'),
mpl_params={'set_xlabel': {'args': ['Heterozygosity'],
'kwargs': {}},
'set_ylabel': {'args': ['SNP number'], 'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}},
'set_yscale': {'args': ['log'], 'kwargs': {}}},
axes=axes, color='c')
axes = fig.add_subplot(212)
title = 'Sample observed Heterozygosity distribution'
plot_distrib(obs_het_sample_distrib, bins=bins2, fhand=open(fpath, 'w'),
mpl_params={'set_xlabel': {'args': ['Heterozygosity'],
'kwargs': {}},
'set_ylabel': {'args': ['Sample number'],
'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}}},
axes=axes, color='c')
canvas.print_figure(fhand)
def plot_maf(variations, data_dir, chunk_size=SNPS_PER_CHUNK, window_size=None,
min_num_genotypes=MIN_NUM_GENOTYPES_FOR_POP_STAT, write_bg=False,
calc_genome_wise=False):
# Calculate and plot MAF distribution
mafs = calc_maf(variations, min_num_genotypes, chunk_size)
maf_distrib, bins = histogram(mafs, n_bins=25, range_=(0, 1))
fpath = join(data_dir, 'mafs.png')
title = 'Maximum allele frequency (MAF) distribution'
plot_distrib(maf_distrib, bins=bins, fhand=open(fpath, 'w'), color='c',
mpl_params={'set_xlabel': {'args': ['MAF'], 'kwargs': {}},
'set_ylabel': {'args': ['SNP number'],
'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}}})
# Write bedgraph file
if calc_genome_wise:
chrom = _load_matrix(variations, CHROM_FIELD)
pos = _load_matrix(variations, POS_FIELD)
bg_fhand = open(join(data_dir, 'maf.bg'), 'w')
pos_maf = PositionalStatsCalculator(chrom, pos, mafs,
window_size=window_size,
step=window_size)
if write_bg:
pos_maf.write(bg_fhand, 'MAF', 'Maximum allele frequency',
track_type='bedgraph')
if window_size is not None:
pos_maf = pos_maf.calc_window_stat()
# Manhattan plot for MAF along genome
fpath = join(data_dir, 'maf_manhattan.png')
fhand = open(fpath, 'w')
title = 'Max Allele Freq (MAF) along the genome'
chrom, pos, mafs = pos_maf.chrom, pos_maf.pos, pos_maf.stat
mpl_params = {'set_xlabel': {'args': ['Chromosome'], 'kwargs': {}},
'set_ylabel': {'args': ['MAF'],'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}}}
manhattan_plot(chrom, pos, mafs, mpl_params=mpl_params,
fhand=fhand, figsize=(15, 7.5))
def plot_r2(variations, window_size, data_dir, write_bg=False):
# Calculate LD r2 parameter in windows
chrom, pos, r2 = calc_r2_windows(variations, window_size=window_size)
# Plot r2 distribution
fpath = os.path.join(data_dir, 'r2_distrib.png')
distrib, bins = histogram(r2, n_bins=50, range_=(0, 1))
title = 'r2 distribution in windows of {} bp'.format(window_size)
mpl_params={'set_xlabel': {'args': ['r2'], 'kwargs': {}},
'set_ylabel': {'args': ['Number of windows'], 'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}}}
plot_distrib(distrib, bins, fhand=open(fpath, 'w'), figsize=(7, 7),
mpl_params=mpl_params)
# Manhattan plot
mask = numpy.logical_not(numpy.isnan(r2))
chrom = chrom[mask]
pos = pos[mask]
r2 = r2[mask]
fpath = os.path.join(data_dir, 'r2_manhattan.png')
title = 'r2 along genome in windows of {} bp'.format(window_size)
mpl_params={'set_xlabel': {'args': ['Chromosome'], 'kwargs': {}},
'set_ylabel': {'args': ['r2'], 'kwargs': {}},
'set_title': {'args': [title], 'kwargs': {}}}
manhattan_plot(chrom, pos, r2, fhand=open(fpath, 'w'), figsize=(15, 7),
marker='k', mpl_params=mpl_params)
# Write bg
if write_bg:
fpath = os.path.join(data_dir, 'r2_windows_{}.png'.format(window_size))
bg_fhand = open(fpath, 'w')
pos_r2 = PositionalStatsCalculator(chrom, pos, r2,
window_size=window_size,
step=window_size,
take_windows=False)
description = 'mean r2 in windows of {} bp'.format(window_size)
pos_r2.write(bg_fhand, 'r2', description, track_type='bedgraph')
def plot_r2(variations, window_size, data_dir, write_bg=False):
# Calculate LD r2 parameter in windows
chrom, pos, r2 = calc_r2_windows(variations, window_size=window_size)
# Plot r2 distribution
fpath = os.path.join(data_dir, 'r2_distrib.png')
distrib, bins = histogram(r2, n_bins=50, range_=(0, 1))
title = 'r2 distribution in windows of {} bp'.format(window_size)
mpl_params = {
'set_xlabel': {
'args': ['r2'],
'kwargs': {}
},
'set_ylabel': {
'args': ['Number of windows'],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
}
}
plot_distrib(distrib,
bins,
fhand=open(fpath, 'w'),
figsize=(7, 7),
mpl_params=mpl_params)
# Manhattan plot
mask = numpy.logical_not(numpy.isnan(r2))
chrom = chrom[mask]
pos = pos[mask]
r2 = r2[mask]
fpath = os.path.join(data_dir, 'r2_manhattan.png')
title = 'r2 along genome in windows of {} bp'.format(window_size)
mpl_params = {
'set_xlabel': {
'args': ['Chromosome'],
'kwargs': {}
},
'set_ylabel': {
'args': ['r2'],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
}
}
manhattan_plot(chrom,
pos,
r2,
fhand=open(fpath, 'w'),
figsize=(15, 7),
marker='k',
mpl_params=mpl_params)
# Write bg
if write_bg:
fpath = os.path.join(data_dir, 'r2_windows_{}.png'.format(window_size))
bg_fhand = open(fpath, 'w')
pos_r2 = PositionalStatsCalculator(chrom,
pos,
r2,
window_size=window_size,
step=window_size,
take_windows=False)
description = 'mean r2 in windows of {} bp'.format(window_size)
pos_r2.write(bg_fhand, 'r2', description, track_type='bedgraph')
def plot_obs_het(variations,
data_dir,
chunk_size=SNPS_PER_CHUNK,
min_num_genotypes=MIN_NUM_GENOTYPES_FOR_POP_STAT):
# Calculate observed heterozygosity distribution by snp
_calc_obs_het_by_var = partial(calc_obs_het,
min_num_genotypes=min_num_genotypes)
distrib = histogram_for_chunks(variations,
calc_funct=_calc_obs_het_by_var,
n_bins=25,
range_=(0, 1),
chunk_size=chunk_size)
obs_het_var_distrib, bins1 = distrib
# Calculate observed heterozygosity distribution by sample
obs_het_by_sample = calc_obs_het_by_sample(variations,
chunk_size=chunk_size)
obs_het_sample_distrib, bins2 = histogram(obs_het_by_sample,
n_bins=25,
range_=(0, 1))
# Plot distributions
fpath = join(data_dir, 'obs_het.png')
fhand = open(fpath, 'w')
fig = Figure(figsize=(10, 10))
canvas = FigureCanvas(fig)
axes = fig.add_subplot(211)
title = 'SNP observed Heterozygosity distribution'
plot_distrib(obs_het_var_distrib,
bins=bins1,
fhand=open(fpath, 'w'),
mpl_params={
'set_xlabel': {
'args': ['Heterozygosity'],
'kwargs': {}
},
'set_ylabel': {
'args': ['SNP number'],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
},
'set_yscale': {
'args': ['log'],
'kwargs': {}
}
},
axes=axes,
color='c')
axes = fig.add_subplot(212)
title = 'Sample observed Heterozygosity distribution'
plot_distrib(obs_het_sample_distrib,
bins=bins2,
fhand=open(fpath, 'w'),
mpl_params={
'set_xlabel': {
'args': ['Heterozygosity'],
'kwargs': {}
},
'set_ylabel': {
'args': ['Sample number'],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
}
},
axes=axes,
color='c')
canvas.print_figure(fhand)
def plot_snp_dens_distrib(variations, window_size, data_dir, write_bg=False):
# Calculate and plot variations density distribution
density = calc_snp_density(variations, window_size)
density_distrib, bins = histogram(density, 20)
fpath = join(data_dir, 'snps_density.png')
title = 'SNP density distribution per {} bp windows'.format(window_size)
plot_distrib(density_distrib,
bins,
fhand=open(fpath, 'w'),
color='c',
mpl_params={
'set_xlabel': {
'args': ['SNP density'],
'kwargs': {}
},
'set_ylabel': {
'args': ['SNP number'],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
},
'set_yscale': {
'args': ['log'],
'kwargs': {}
}
})
# Manhattan plot for SNP density
fpath = join(data_dir, 'snps_density_manhattan.png')
fhand = open(fpath, 'w')
title = 'SNP denisity along the genome'
chrom = _load_matrix(variations, CHROM_FIELD)
pos = _load_matrix(variations, POS_FIELD)
manhattan_plot(chrom,
pos,
density,
mpl_params={
'set_xlabel': {
'args': ['Chromosome'],
'kwargs': {}
},
'set_ylabel': {
'args': ['SNP per {} bp'.format(window_size)],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
}
},
fhand=fhand,
figsize=(15, 7.5),
ylim=1)
# Save in bedgraph format
if write_bg:
bg_fhand = open(join(data_dir, 'snp_density.bg'), 'w')
pos_dens = PositionalStatsCalculator(chrom, pos, density)
pos_dens.write(bg_fhand,
'snp_density',
'SNP number in {} bp around'.format(window_size),
track_type='bedgraph')
def plot_call_field_distribs_per_gt_type(variations,
field,
max_value,
data_dir,
chunk_size=SNPS_PER_CHUNK):
# Field distribution per sample
field_name = field.split('/')[-1]
fpath = join(data_dir, '{}_distribution_per_sample.png'.format(field_name))
mask_funcs = [call_is_het, call_is_hom]
names = ['Heterozygous', 'Homozygous']
distribs = []
for mask_func in mask_funcs:
dp_distribs, bins = calc_field_distribs_per_sample(
variations,
field=field,
range_=(0, max_value),
n_bins=max_value,
chunk_size=chunk_size,
mask_func=mask_func,
mask_field=GT_FIELD)
distribs.append(dp_distribs)
title = '{} distribution per sample'.format(field_name)
mpl_params = {
'set_xlabel': {
'args': ['Samples'],
'kwargs': {}
},
'set_ylabel': {
'args': [field_name],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
}
}
figsize = (variations[GT_FIELD].shape[1], 7)
plot_boxplot_from_distribs_series(distribs,
fhand=open(fpath, 'w'),
mpl_params=mpl_params,
figsize=figsize,
colors=['pink', 'tan'],
labels=names,
xticklabels=variations.samples)
# Overall field distributions
fpath = join(data_dir, '{}_distribution.png'.format(field_name))
fhand = open(fpath, 'w')
fig = Figure(figsize=(20, 15))
canvas = FigureCanvas(fig)
i = 1
for distrib, name in zip(distribs, names):
distrib = numpy.sum(dp_distribs, axis=0)
distrib_cum = calc_cum_distrib(distrib)
axes = fig.add_subplot(len(names) * 100 + 20 + i)
i += 1
title = '{} distribution all samples {}'.format(field_name, name)
plot_distrib(distrib,
bins,
axes=axes,
mpl_params={
'set_xlabel': {
'args': [field_name],
'kwargs': {}
},
'set_ylabel': {
'args': ['Number of GTs'],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
}
})
distrib_cum = distrib_cum / distrib_cum[0] * 100
axes = fig.add_subplot(len(names) * 100 + 20 + i)
i += 1
title = '{} cumulative distribution all samples {}'.format(
field_name, name)
plot_distrib(distrib_cum,
bins,
axes=axes,
mpl_params={
'set_xlabel': {
'args': [field_name],
'kwargs': {}
},
'set_ylabel': {
'args': ['% calls > Depth '],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
}
})
canvas.print_figure(fhand)
def plot_inbreeding_coefficient(
variations,
max_num_allele,
data_dir,
window_size,
chunk_size=SNPS_PER_CHUNK,
min_num_genotypes=MIN_NUM_GENOTYPES_FOR_POP_STAT,
write_bg=False,
calc_genome_wise=False):
# Calculate Inbreeding coefficient distribution
inbreed_coef = calc_inbreeding_coef(variations,
chunk_size=chunk_size,
min_num_genotypes=min_num_genotypes)
ic_distrib, bins = histogram(inbreed_coef, 50, range_=(-1, 1))
fpath = join(data_dir, 'inbreeding_coef_distribution.png')
fhand = open(fpath, 'w')
title = 'Inbreeding coefficient distribution all samples'
plot_distrib(ic_distrib,
bins,
fhand=fhand,
mpl_params={
'set_xlabel': {
'args': ['Inbreeding coefficient'],
'kwargs': {}
},
'set_ylabel': {
'args': ['Number of SNPs'],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
},
'set_xlim': {
'args': [-1, 1],
'kwargs': {}
}
})
# Save in bedgraph file
if calc_genome_wise:
bg_fhand = open(join(data_dir, 'ic.bg'), 'w')
chrom = _load_matrix(variations, CHROM_FIELD)
pos = _load_matrix(variations, POS_FIELD)
pos_ic = PositionalStatsCalculator(chrom, pos, inbreed_coef)
if write_bg:
pos_ic.write(bg_fhand,
'IC',
'Inbreeding coefficient',
track_type='bedgraph')
# Plot Ic along genome taking sliding windows
pos_ic = pos_ic.calc_window_stat()
chrom, pos, ic_windows = pos_ic.chrom, pos_ic.pos, pos_ic.stat
fpath = join(data_dir, 'ic_manhattan.png')
fhand = open(fpath, 'w')
title = 'Inbreeding coefficient (IC) along the genome'
manhattan_plot(chrom,
pos,
ic_windows,
fhand=fhand,
figsize=(15, 7.5),
ylim=-1,
mpl_params={
'set_xlabel': {
'args': ['Chromosome'],
'kwargs': {}
},
'set_ylabel': {
'args': ['IC'],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
}
})
def plot_maf(variations,
data_dir,
chunk_size=SNPS_PER_CHUNK,
window_size=None,
min_num_genotypes=MIN_NUM_GENOTYPES_FOR_POP_STAT,
write_bg=False,
calc_genome_wise=False):
# Calculate and plot MAF distribution
mafs = calc_maf(variations, min_num_genotypes, chunk_size)
maf_distrib, bins = histogram(mafs, n_bins=25, range_=(0, 1))
fpath = join(data_dir, 'mafs.png')
title = 'Maximum allele frequency (MAF) distribution'
plot_distrib(maf_distrib,
bins=bins,
fhand=open(fpath, 'w'),
color='c',
mpl_params={
'set_xlabel': {
'args': ['MAF'],
'kwargs': {}
},
'set_ylabel': {
'args': ['SNP number'],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
}
})
# Write bedgraph file
if calc_genome_wise:
chrom = _load_matrix(variations, CHROM_FIELD)
pos = _load_matrix(variations, POS_FIELD)
bg_fhand = open(join(data_dir, 'maf.bg'), 'w')
pos_maf = PositionalStatsCalculator(chrom,
pos,
mafs,
window_size=window_size,
step=window_size)
if write_bg:
pos_maf.write(bg_fhand,
'MAF',
'Maximum allele frequency',
track_type='bedgraph')
if window_size is not None:
pos_maf = pos_maf.calc_window_stat()
# Manhattan plot for MAF along genome
fpath = join(data_dir, 'maf_manhattan.png')
fhand = open(fpath, 'w')
title = 'Max Allele Freq (MAF) along the genome'
chrom, pos, mafs = pos_maf.chrom, pos_maf.pos, pos_maf.stat
mpl_params = {
'set_xlabel': {
'args': ['Chromosome'],
'kwargs': {}
},
'set_ylabel': {
'args': ['MAF'],
'kwargs': {}
},
'set_title': {
'args': [title],
'kwargs': {}
}
}
manhattan_plot(chrom,
pos,
mafs,
mpl_params=mpl_params,
fhand=fhand,
figsize=(15, 7.5))
