def plot_idr_vs_peaks(self, filename=None, savefig=False):
# global_idr is actually -log10(idr)
pylab.clf()
X1 = pylab.linspace(0, self.threshold, 100)
X2 = pylab.linspace(self.threshold, 1, 100)
# convert global idr to proba
df1 = self.df.query("idr<@self.threshold")
df2 = self.df.query("idr>[email protected]")
pylab.plot([sum(df1['idr'] < x) for x in X1], X1, '-', color='r', lw=2)
shift = len(df1)
pylab.plot([shift + sum(df2['idr'] < x) for x in X2],
X2,
"-",
color='k',
lw=2)
pylab.xlabel('Number of significant peaks')
pylab.ylabel('IDR')
pylab.axhline(0.05, color='b', ls='--')
pylab.axvline(self.N_significant_peaks, color='b', ls='--')
if savefig:
pylab.savefig(filename)
def get_percentage_genes_covered_at_this_fraction(self, this):
assert this <= 1 and this >= 0
icol = self.coverage_column
X = pylab.linspace(0, 1, 101)
N = float(len(self.df))
Y = np.array([sum(self.df[icol] > x) / N * 100 for x in X])
return np.interp(this, X, Y)
def plot_gc_content(self, fontsize=16, ec="k", bins=100):
"""plot GC content histogram
:params bins: a value for the number of bins or an array (with a copy()
method)
:param ec: add black contour on the bars
.. plot::
:include-source:
from sequana import BAM, sequana_data
b = BAM(sequana_data('test.bam'))
b.plot_gc_content()
"""
data = self.get_gc_content()
try:
X = np.linspace(0, 100, bins)
except:
X = bins.copy()
pylab.hist(data, X, normed=True, ec=ec)
pylab.grid(True)
mu = pylab.mean(data)
sigma = pylab.std(data)
X = pylab.linspace(X.min(), X.max(), 100)
pylab.plot(X, pylab.normpdf(X, mu, sigma), lw=2, color="r", ls="--")
pylab.xlabel("GC content", fontsize=16)
def plot_gc_content(self, fontsize=16, ec="k", bins=100):
"""plot GC content histogram
:params bins: a value for the number of bins or an array (with a copy()
method)
:param ec: add black contour on the bars
.. plot::
:include-source:
from sequana import BAM, sequana_data
b = BAM(sequana_data('test.bam'))
b.plot_gc_content()
"""
data = self.get_gc_content()
try:
X = np.linspace(0, 100, bins)
except:
X = bins.copy()
pylab.hist(data, X, density=True, ec=ec)
pylab.grid(True)
mu = pylab.mean(data)
sigma = pylab.std(data)
X = pylab.linspace(X.min(), X.max(), 100)
from sequana.misc import normpdf
pylab.plot(X, normpdf(X, mu, sigma), lw=2, color="r", ls="--")
pylab.xlabel("GC content", fontsize=16)
def hist_snr(self, bins=50, alpha=0.5, hold=False, fontsize=12,
grid=True, xlabel="SNR", ylabel="#",title="", clip_upper_SNR=30):
"""Plot histogram of the ACGT SNRs for all reads
:param int bins: binning for the histogram. Note that the range starts
at 0 and ends at clip_upper_SNR
:param float alpha: transparency of the histograms
:param bool hold:
:param int fontsize:
:param bool grid:
:param str xlabel:
:param str ylabel:
:param str title:
.. plot::
:include-source:
from sequana.pacbio import PacbioSubreads
from sequana import sequana_data
b = PacbioSubreads(sequana_data("test_pacbio_subreads.bam"))
b.hist_snr()
"""
if self._df is None:
self._get_df()
# old pacbio format has no SNR stored
if len(self._df['snr_A'].dropna()) == 0:
# nothing to plot
from sequana import sequana_data
pylab.clf()
pylab.imshow(pylab.imread(sequana_data("no_data.jpg")))
pylab.gca().axis('off')
return
if hold is False:
pylab.clf()
maxSNR = 0
for letter in "ACGT":
m = self._df.loc[:,"snr_{}".format(letter)].max()
if m > maxSNR:
maxSNR = m
if maxSNR > clip_upper_SNR:
maxSNR = clip_upper_SNR
bins = pylab.linspace(0, maxSNR, bins)
pylab.hist(self._df.loc[:,'snr_A'].clip_upper(maxSNR), alpha=alpha, label="A", bins=bins)
pylab.hist(self._df.loc[:,'snr_C'].clip_upper(maxSNR), alpha=alpha, label="C", bins=bins)
pylab.hist(self._df.loc[:,'snr_G'].clip_upper(maxSNR), alpha=alpha, label="G", bins=bins)
pylab.hist(self._df.loc[:,'snr_T'].clip_upper(maxSNR), alpha=alpha, label="T", bins=bins)
pylab.legend()
pylab.xlabel(xlabel, fontsize=fontsize)
pylab.ylabel(ylabel, fontsize=fontsize)
pylab.title(title,fontsize=fontsize)
if grid is True:
pylab.grid(True)
def plot(self,
normed=True,
N=1000,
Xmin=None,
Xmax=None,
bins=50,
color='red',
lw=2,
hist_kw={
'color': '#5F9EA0',
"edgecolor": "k"
},
ax=None):
if ax:
ax.hist(self.data, normed=normed, bins=bins, **hist_kw)
else:
pylab.hist(self.data, density=normed, bins=bins, **hist_kw)
if Xmin is None:
Xmin = self.data.min()
if Xmax is None:
Xmax = self.data.max()
X = pylab.linspace(Xmin, Xmax, N)
if ax:
ax.plot(X, [self.model.pdf(x, self.results.x) for x in X],
color=color,
lw=lw)
else:
pylab.plot(X, [self.model.pdf(x, self.results.x) for x in X],
color=color,
lw=lw)
K = len(self.results.x)
# The PIs must be normalised
import scipy.stats as ss
for i in range(self.k):
mu, sigma, pi_ = self.results.mus[i], self.results.sigmas[
i], self.results.pis[i]
if ax:
ax.plot(X, [pi_ * ss.norm.pdf(x, mu, sigma) for x in X],
'k--',
alpha=0.7,
lw=2)
else:
pylab.plot(X, [pi_ * ss.norm.pdf(x, mu, sigma) for x in X],
'k--',
alpha=0.7,
lw=2)
def hist_snr(self,
bins=50,
alpha=0.5,
hold=False,
fontsize=12,
grid=True,
xlabel="SNR",
ylabel="#",
title="",
clip_upper_SNR=30):
"""Plot histogram of the ACGT SNRs for all reads
:param int bins: binning for the histogram. Note that the range starts
at 0 and ends at clip_upper_SNR
:param float alpha: transparency of the histograms
:param bool hold:
:param int fontsize:
:param bool grid:
:param str xlabel:
:param str ylabel:
:param str title:
.. plot::
:include-source:
from sequana.pacbio import PacbioSubreads
from sequana import sequana_data
b = PacbioSubreads(sequana_data("test_pacbio_subreads.bam"))
b.hist_snr()
"""
if self._df is None:
self._get_df()
# old pacbio format has no SNR stored
if len(self._df['snr_A'].dropna()) == 0:
# nothing to plot
from sequana import sequana_data
pylab.clf()
pylab.imshow(pylab.imread(sequana_data("no_data.jpg")))
pylab.gca().axis('off')
return
if hold is False:
pylab.clf()
maxSNR = 0
for letter in "ACGT":
m = self._df.loc[:, "snr_{}".format(letter)].max()
if m > maxSNR:
maxSNR = m
if maxSNR > clip_upper_SNR:
maxSNR = clip_upper_SNR
bins = pylab.linspace(0, maxSNR, bins)
pylab.hist(self._df.loc[:, 'snr_A'].clip_upper(maxSNR),
alpha=alpha,
label="A",
bins=bins)
pylab.hist(self._df.loc[:, 'snr_C'].clip_upper(maxSNR),
alpha=alpha,
label="C",
bins=bins)
pylab.hist(self._df.loc[:, 'snr_G'].clip_upper(maxSNR),
alpha=alpha,
label="G",
bins=bins)
pylab.hist(self._df.loc[:, 'snr_T'].clip_upper(maxSNR),
alpha=alpha,
label="T",
bins=bins)
pylab.legend()
pylab.xlabel(xlabel, fontsize=fontsize)
pylab.ylabel(ylabel, fontsize=fontsize)
pylab.title(title, fontsize=fontsize)
if grid is True:
pylab.grid(True)
def plot_jaccard_distance(self, mode, padjs=[0.0001,0.001,0.01,0.05,0.1],
Nfc=50, smooth=False, window=5):
assert mode in ['down', 'up', 'all']
pylab.clf()
if mode == "down":
m1 = self.r1.df.log2FoldChange.min()
m2 = self.r2.df.log2FoldChange.min()
minimum = min(m1,m2)
print(m1, m2)
X = pylab.linspace(0, minimum, Nfc)
elif mode == "up":
m1 = self.r1.df.log2FoldChange.max()
m2 = self.r2.df.log2FoldChange.max()
maximum = max(m1,m2)
X = pylab.linspace(0, maximum, Nfc)
else:
minmax1 = self.r1.df.log2FoldChange.abs().max()
minmax2 = self.r2.df.log2FoldChange.abs().max()
maximum = max(minmax1, minmax2)
X = pylab.linspace(0, maximum, Nfc)
common = {}
for padj in padjs:
I = []
common[padj] = []
for x in X:
if mode == "down":
# less than a given fold change that is negative
A = set(self.r1.df.query("log2FoldChange<=@x and padj<@padj").index)
B = set(self.r2.df.query("log2FoldChange<=@x and padj<@padj").index)
elif mode == "up":
# greater than a given fold change that is positive
A = set(self.r1.df.query("log2FoldChange>=@x and padj<@padj").index)
B = set(self.r2.df.query("log2FoldChange>=@x and padj<@padj").index)
else:
A = set(self.r1.df.query("(log2FoldChange>=@x or log2FoldChange<=-@x) and padj<@padj").index)
B = set(self.r2.df.query("(log2FoldChange>=@x or log2FoldChange<=-@x) and padj<@padj").index)
if len(A) == 0 or len(B) == 0:
# no overlap yet
I.append(100)
else:
res = len(A.intersection(B)) / (len(A) + len(B) - len(A.intersection(B))) * 100
I.append(res)
common[padj].append(len(A.intersection(B)))
try:
if smooth:
I = pd.Series(I).rolling(window).median().values
else:
assert False
except:
pass
pylab.plot(X, I, 'o-', label=str(padj))
ax = pylab.gca()
ax.set_ylabel("Jaccard similarity (intersection/union)")
ax.set_xlabel("Fold change (log2)")
ax2 = ax.twinx()
for padj in padjs:
ax2.plot(X, common[padj], color='orange', ls='--')
ax2.set_ylabel("Cardinality of the union ")
ax.legend()
ax.set_ylim([0,100])
#ax2.set_ylim([0,100])
if mode == "down":
ax.axvline(-2, ls='--', color='r')
else:
ax.axvline(2, ls='--', color='r')