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 hist_read_length_consensus_isoform(self, mode="all", bins=80, rwidth=0.8,
align="left", fontsize=16, edgecolor="k", **kwargs):
"""
mode can be all, lq, hq
"""
pylab.clf()
L1 = [len(read['sequence']) for read in self.lq_sequence]
L2 = [len(read['sequence']) for read in self.hq_sequence]
if mode == "all":
L = L1 + L2
elif mode == "lq":
L = L1
else:
L = L2
Y, X, _ = pylab.hist(L, bins=bins, rwidth=rwidth, align=align,
ec=edgecolor, **kwargs)
pylab.gca().set_ylim(bottom=0)
pylab.gca().set_xlim(left=0)
pylab.xlabel("Read length", fontsize=fontsize)
pylab.ylabel("Number of reads", fontsize=fontsize)
pylab.grid()
ax_twin = pylab.gca().twinx()
shift = (X[1] - X[0]) / 2
ax_twin.plot(X[0:-1]- shift, len(L) - pylab.cumsum(Y), "k")
ax_twin.set_ylim(bottom=0)
pylab.ylabel("CDF", fontsize=fontsize)
pylab.title("Read length of Consensus isoforms reads")
def scatter_plot(self, filename=None, hold=False):
"""Scatter plot of the score versus length of each ortholog
.. plot::
:include-source:
from sequana import BUSCO, sequana_data
b = BUSCO(sequana_data("test_busco_full_table.tsv"))
b.scatter_plot()
"""
if hold is False:
pylab.clf()
colors = ["green", "orange", "red", "blue"]
markers = ['o', 's', 'x', 'o']
for i, this in enumerate(["Complete", "Fragmented", "Missing", "Duplicated"]):
mask = self.df.Status == "Complete"
if sum(mask)>0:
self.df[mask].plot(x="Length", y="Score", kind="scatter",
color=colors[i],
marker=markers[i], label="Complete")
pylab.legend()
pylab.grid()
if filename:
pylab.savefig(filename)
def hist_concordance(self, bins=100, fontsize=16):
"""
formula : 1 - (in + del + mismatch / (in + del + mismatch + match) )
For BWA and BLASR, the get_cigar_stats are different !!!
BWA for instance has no X stored while Pacbio forbids the use of the M
(CMATCH) tag. Instead, it uses X (CDIFF) and = (CEQUAL) characters.
Subread Accuracy: The post-mapping accuracy of the basecalls.
Formula: [1 - (errors/subread length)], where errors = number of deletions +
insertions + substitutions.
"""
try:
concordance = self._concordance
except:
self._set_concordance()
concordance = self._concordance
pylab.hist(concordance, bins=bins)
pylab.grid()
mu = np.mean(concordance)
median = np.median(concordance)
pylab.axvline(mu, color='r', alpha=0.5)
pylab.axvline(median, color='r', alpha=0.5, ls="--")
pylab.xlabel("concordance", fontsize=fontsize)
def plot_count_per_sample(self, fontsize=12, sample_list=None):
""""Number of mapped reads per sample. Each color for each replicate
.. plot::
:include-source:
from sequana.rnadiff import RNADiffResults
from sequana import sequana_data
r = RNADiffResults(sequana_data("rnadiff/rnadiff_onecond_1"))
r.plot_count_per_sample()
"""
sample_names = self.sample_names
N = len(sample_names)
dd = self.df[sample_names].sum()
pylab.clf()
colors = []
for sample in self.sample_names:
colors.append(self.colors[self.get_cond_from_sample(sample)])
pylab.bar(range(N), (dd/1000000).values,
color=colors, alpha=1,
zorder=10, lw=1, ec="k", width=0.9)
pylab.xlabel("Samples", fontsize=fontsize)
pylab.ylabel("Total read count (millions)", fontsize=fontsize)
pylab.grid(True, zorder=0)
pylab.title("Total read count per sample", fontsize=fontsize)
pylab.xticks(range(N), self.sample_names)
def plot_volcano(self):
"""
.. plot::
:include-source:
from sequana.rnadiff import RNADiffResults
from sequana import sequana_data
r = RNADiffResults(sequana_data("rnadiff/rnadiff_onecond_1"))
r.plot_volcano()
"""
d1 = self.df.query("padj>0.05")
d2 = self.df.query("padj<=0.05")
fig = pylab.figure()
pylab.plot(d1.log2FoldChange, -np.log10(d1.padj), marker="o",
alpha=0.5, color="r", lw=0)
pylab.plot(d2.log2FoldChange, -np.log10(d2.padj), marker="o",
alpha=0.5, color="k", lw=0)
pylab.grid(True)
pylab.xlabel("fold change")
pylab.ylabel("log10 adjusted p-value")
m1 = abs(min(self.df.log2FoldChange))
m2 = max(self.df.log2FoldChange)
limit = max(m1,m2)
pylab.xlim([-limit, limit])
y1,y2 = pylab.ylim()
pylab.ylim([0,y2])
pylab.axhline(-np.log10(0.05), lw=2, ls="--", color="r", label="pvalue threshold (0.05)")
def plot_hist_normalized_coverage(self, filename=None, binwidth=0.1,
max_z=4):
""" Barplot of the normalized coverage with gaussian fitting
"""
pylab.clf()
# if there are a NaN -> can't set up binning
d = self.df["scale"][self.range[0]:self.range[1]].dropna()
# remove outlier -> plot crash if range between min and max is too high
d = d[np.abs(d - d.mean()) <= (4 * d.std())]
bins = self._set_bins(d, binwidth)
self.mixture_fitting.data = d
try:
self.mixture_fitting.plot(self.gaussians_params, bins=bins, Xmin=0,
Xmax=max_z)
except ZeroDivisionError:
pass
pylab.grid(True)
pylab.xlim([0,max_z])
pylab.xlabel("Normalised per-base coverage")
try:
pylab.tight_layout()
except:
pass
if filename:
pylab.savefig(filename)
def plot_unknown_barcodes(self, N=20):
ub = self.data['UnknownBarcodes']
df = pd.DataFrame({x['Lane']: x['Barcodes'] for x in ub})
if "unknown" in df.index and len(df) == 1:
df.loc['known'] = [0 for i in df.columns]
# if data is made of undetermined only, the dataframe is just made of
# N lanes with one entry : unknown
S = df.sum(axis=1).sort_values(ascending=False).index[0:N]
data = df.loc[S][::-1]
#print(data)
data.columns = ["Lane {}".format(x) for x in data.columns]
from matplotlib import rcParams
rcParams['axes.axisbelow'] = True
pylab.figure(figsize=(10, 8))
ax = pylab.gca()
data.plot(kind="barh", width=1, ec="k", ax=ax)
rcParams['axes.axisbelow'] = False
pylab.xlabel("Number of reads", fontsize=12)
pylab.ylabel("")
pylab.grid(True)
pylab.legend(
["Lane {}".format(x) for x in range(1,
len(df.columns) + 1)],
loc="lower right")
try:
pylab.tight_layout()
except Exception as err:
print(err)
return data
def scatterplot(self, enrich, cutoff=0.05, nmax=10, gene_set_size=[]):
df = self._get_final_df(enrich.results, cutoff=cutoff, nmax=nmax)
pylab.clf()
pylab.scatter(-pylab.log10(df['Adjusted P-value']),
range(len(df)),
s=10 * df['size'],
c=df['size'])
pylab.xlabel("Odd ratio")
pylab.ylabel("Gene sets")
pylab.yticks(range(len(df)), df.name)
a, b = pylab.xlim()
pylab.xlim([0, b])
pylab.grid(True)
ax = pylab.gca()
M = max(df['size'])
if M > 100:
l1, l2, l3 = "10", "100", str(M)
else:
l1, l2, l3 = str(round(M / 3)), str(round(M * 2 / 3)), str(M)
handles = [
pylab.Line2D([0], [0], marker="o", markersize=5, label=l1, ls=""),
pylab.Line2D([0], [0], marker="o", markersize=10, label=l2, ls=""),
pylab.Line2D([0], [0], marker="o", markersize=15, label=l3, ls="")
]
ax.legend(handles=handles, loc="upper left", title="gene-set size")
pylab.axvline(1.3, lw=2, ls="--", color="r")
pylab.tight_layout()
ax = pylab.colorbar(pylab.gci())
return df
def get_max_gc_correlation(self, reference):
"""Plot correlation between coverage and GC content by varying the GC window
The GC content uses a moving window of size W. This parameter affects
the correlation bewteen coverage and GC. This function find the
*optimal* window length.
"""
pylab.clf()
corrs = []
wss = []
def func(params):
ws = int(round(params[0]))
if ws < 10:
return 0
self.bed.compute_gc_content(reference, ws)
corr = self.get_gc_correlation()
corrs.append(corr)
wss.append(ws)
return corr
from scipy.optimize import fmin
res = fmin(func, 100, xtol=1, disp=False) # guess is 200
pylab.plot(wss, corrs, "o")
pylab.xlabel("GC window size")
pylab.ylabel("Correlation")
pylab.grid()
return res[0]
def histogram_sequence_lengths(self, logy=True):
"""Histogram sequence lengths
.. plot::
:include-source:
from sequana import sequana_data
from sequana import FastQC
filename = sequana_data("test.fastq", "testing")
qc = FastQC(filename)
qc.histogram_sequence_lengths()
"""
data = [len(x) for x in self.sequences]
bary, barx = np.histogram(data, bins=range(max(data)+1))
# get rid of zeros to avoid warnings
bx = [x for x,y in zip(barx, bary) if y!=0]
by = [y for x,y in zip(barx, bary) if y!=0]
if logy:
pylab.bar(bx, pylab.log10(by))
else:
pylab.bar(bx, by)
pylab.xlim([1,max(data)+1])
pylab.grid(True)
pylab.xlabel("position (bp)", fontsize=self.fontsize)
pylab.ylabel("Count (log scale)", fontsize=self.fontsize)
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 plotter(filename, key):
name = key.replace(" ", "_")
pylab.ioff()
histograms[key].plot(logy=False, lw=2, marker="o")
pylab.title(name + "(%s)" % count)
pylab.grid(True)
pylab.savefig(filename)
pylab.close() # need to close the figure otherwise warnings
def plot_padj_hist(self, bins=60, fontsize=16):
pylab.hist(self.df.padj.dropna(), bins=bins, ec="k")
pylab.grid(True)
pylab.xlabel("Adjusted p-value", fontsize=fontsize)
pylab.ylabel("Occurences", fontsize=fontsize)
try:
pylab.tight_layout()
except:
pass
def plot_pvalue_hist(self, bins=60, fontsize=16, rotation=0):
pylab.hist(self.df.pvalue.dropna(), bins=bins, ec="k")
pylab.grid(True)
pylab.xlabel("raw p-value", fontsize=fontsize)
pylab.ylabel("Occurences", fontsize=fontsize)
try:
pylab.tight_layout()
except:
pass
def barplot_per_sample(self, alpha=0.5, width=0.8, filename=None):
df = self.get_data_reads()
# this is ugly but will do the job for now
under = df.query("name=='Undetermined'")
others = df.query("name!='Undetermined'")
under = under.groupby("name").sum().reset_index()
others = others.groupby("name").sum().reset_index()
under = under[["name", "count"]].set_index("name")
others = others[["name", "count"]].set_index("name")
all_data = others.sort_index(ascending=False)
all_data.columns = ["samples"]
# appended at the end
all_data.loc['undetermined'] = 0
# revert back
all_data = all_data.loc[::-1]
# just for legend
under.columns = ['undetermined']
if all_data.sum().min() > 1e6:
all_data /= 1e6
under /= 1e6
M = True
else:
M = False
all_data.plot(kind="barh", alpha=alpha, zorder=1, width=width, ec='k')
under.plot(kind="barh",
alpha=alpha,
color="red",
ax=pylab.gca(),
zorder=1,
width=width,
ec='k')
pylab.ylim([-0.5, len(all_data) + 0.5])
if len(all_data) < 100:
pylab.yticks(range(len(all_data)), all_data.index)
pylab.legend()
pylab.grid(True, zorder=-1)
if M:
pylab.xlabel("Number of reads (M)")
else:
pylab.xlabel("Number of reads")
try:
pylab.tight_layout()
except:
pass
if filename:
pylab.savefig(filename, dpi=200)
def plot_volcano_differences(self, mode="all"):
cond1, cond2 = "cond1", "cond2"
labels = [cond1, cond2]
A = self.r1.df.loc[self.r1.gene_lists[mode]]
B = self.r2.df.loc[self.r2.gene_lists[mode]]
AB = set(A.index).intersection(set(B.index))
Aonly = A.loc[set(A.index).difference(set(B.index))]
Bonly = B.loc[set(B.index).difference(set(A.index))]
Acommon = A.loc[AB]
Bcommon = B.loc[AB]
pylab.clf()
pylab.plot(Acommon.log2FoldChange, -np.log10(Acommon.padj), marker="o",
alpha=0.5, color="r", lw=0, label="Common in experiment 1", pickradius=4,
picker=True)
pylab.plot(Bcommon.log2FoldChange, -np.log10(Bcommon.padj), marker="o",
alpha=0.5, color="orange", lw=0, label="Common in experiment 2", pickradius=4,
picker=True)
for x in AB:
a_l = A.loc[x].log2FoldChange
a_p = -np.log10(A.loc[x].padj)
b_l = B.loc[x].log2FoldChange
b_p = -np.log10(B.loc[x].padj)
pylab.plot([a_l, b_l], [a_p, b_p], 'k', alpha=0.5)
pylab.plot(Bonly.log2FoldChange, -np.log10(Bonly.padj), marker="*",
alpha=0.5, color="blue", lw=0, label="In experiment 2 only", pickradius=4,
picker=True)
pylab.plot(Aonly.log2FoldChange, -np.log10(Aonly.padj), marker="*",
alpha=0.5, color="cyan", lw=0, label="In experiment 1 only", pickradius=4,
picker=True)
for name, x in Bonly.iterrows():
x1 = x.log2FoldChange
y1 = -np.log10(x.padj)
x2 = self.r1.df.loc[name].log2FoldChange
y2 = -np.log10(self.r1.df.loc[name].padj)
pylab.plot( [x1,x2], [y1,y2], ls="--", color='r')
for name, x in Aonly.iterrows():
x1 = x.log2FoldChange
y1 = -np.log10(x.padj)
x2 = self.r2.df.loc[name].log2FoldChange
y2 = -np.log10(self.r2.df.loc[name].padj)
pylab.plot( [x1,x2], [y1,y2], ls="-", color='r')
pylab.axhline(1.33, alpha=0.5, ls="--", color="r")
pylab.xlabel("log2 fold Change")
pylab.ylabel("log10 adjusted p-values")
pylab.legend()
pylab.grid(True)
return Aonly, Bonly, Acommon, Bcommon
def plot(self, fontsize=16):
"""plot quality versus base position"""
pylab.plot(self.quality, label="offset: %s" % self.offset)
pylab.xlabel('base position', fontsize=fontsize)
pylab.ylabel('Quality per base', fontsize=fontsize)
pylab.grid(True)
# ylim set autoscale to off so if we want to call this function several
# times, we must reset autoscale to on before calling ylim
pylab.autoscale()
limits = pylab.ylim()
pylab.ylim(max(0,limits[0]-1), limits[1]+1)
def plot(self, fontsize=16):
"""plot quality versus base position"""
pylab.plot(self.quality, label="offset: %s" % self.offset)
pylab.xlabel('base position', fontsize=fontsize)
pylab.ylabel('Quality per base', fontsize=fontsize)
pylab.grid(True)
# ylim set autoscale to off so if we want to call this function several
# times, we must reset autoscale to on before calling ylim
pylab.autoscale()
limits = pylab.ylim()
pylab.ylim(max(0, limits[0] - 1), limits[1] + 1)
def boxplot_mapq_concordance(self):
# method can only be bwa for now
assert self.method == "bwa"
data = self._get_data()
df = pd.DataFrame(data, columns=["mapq", "length", "concordance"])
pylab.clf()
pylab.boxplot([df[df.mapq == i]['concordance'] for i in range(1,61)])
pylab.xlabel("mapq")
pylab.ylabel("concordance")
pylab.grid()
tt = [10,20,30,40,50,60]
pylab.xticks(tt, tt)
def plot_genesets_hist(self, bins=20):
N = len(self.gene_sets.keys())
pylab.clf()
pylab.hist([len(v) for k, v in self.gene_sets.items()],
bins=bins,
lw=1,
ec="k")
pylab.title("{} gene sets".format(N))
pylab.xlabel("Gene set sizes")
pylab.grid(True)
a, b = pylab.xlim()
pylab.xlim([0, b])
def plot_indel_dist(self, fontsize=16):
"""Plot indel count (+ ratio)
:Return: list of insertions, deletions and ratio insertion/deletion for
different length starting at 1
.. plot::
:include-source:
from sequana import sequana_data, BAM
b = BAM(sequana_data("measles.fa.sorted.bam"))
b.plot_indel_dist()
What you see on this figure is the presence of 10 insertions of length
1, 1 insertion of length 2 and 3 deletions of length 1
# Note that in samtools, several insertions or deletions in a single
alignment are ignored and only the first one seems to be reported. For
instance 10M1I10M1I stored only 1 insertion in its report; Same comment
for deletions.
.. todo:: speed up and handle long reads cases more effitiently by
storing INDELS as histograms rather than lists
"""
try:
self.insertions
except:
self._set_indels()
if len(self.insertions) ==0 or len(self.deletions) == 0:
raise ValueError("No deletions or insertions found")
N = max(max(Counter(self.deletions)), max(Counter(self.insertions))) + 1
D = [self.deletions.count(i) for i in range(N)]
I = [self.insertions.count(i) for i in range(N)]
R = [i/d if d!=0 else 0 for i,d in zip(I, D)]
fig, ax = pylab.subplots()
ax.plot(range(N), I, marker="x", label="Insertions")
ax.plot(range(N), D, marker="x", label="Deletions")
ax.plot(range(N), R, "--r", label="Ratio insertions/deletions")
ax.set_yscale("symlog")
pylab.ylim([1, pylab.ylim()[1]])
pylab.legend()
pylab.grid()
from matplotlib.ticker import MaxNLocator
ax.xaxis.set_major_locator(MaxNLocator(integer=True))
pylab.xlabel("Indel length", fontsize=fontsize)
pylab.ylabel("Indel count", fontsize=fontsize)
return I, D, R
def hist_coverage(self, bins=100):
"""
.. plot::
:include-source:
from sequana import sequana_data, BAM
b = BAM(sequana_data("measles.fa.sorted.bam"))
b.hist_coverage()
"""
try: self.coverage
except: self._set_coverage()
pylab.hist(self.coverage, bins=bins)
pylab.xlabel("Coverage")
pylab.ylabel("Number of mapped bases")
pylab.grid()
def hist_length_repeats(self, bins=20, alpha=0.5, hold=False,
fontsize=12, grid=True, title="Repeat length",
xlabel="Repeat length", ylabel="#", logy=True):
"""Plots histogram of the repeat lengths
"""
# check that user has set a threshold
if hold is False:
pylab.clf()
pylab.hist(self.list_len_repeats, alpha=alpha, bins=bins)
pylab.title(title)
pylab.xlabel(xlabel, fontsize=fontsize)
pylab.ylabel(ylabel, fontsize=fontsize)
if grid is True:
pylab.grid(True)
if logy:
pylab.semilogy()
def plot_read_length(self):
"""Plot occurences of aligned read lengths
.. plot::
:include-source:
from sequana import sequana_data, BAM
b = BAM(sequana_data("test.bam"))
b.plot_read_length()
"""
X, Y = self._get_read_length()
pylab.plot(X, Y,
label="min length:{}; max length:{}".format(min(X), max(X)))
pylab.grid()
pylab.xlabel("Read length", fontsize=16)
pylab.legend()
def histogram_gc_content(self):
"""Plot histogram of GC content
.. plot::
:include-source:
from sequana import sequana_data
from sequana import FastQC
filename = sequana_data("test.fastq", "testing")
qc = FastQC(filename)
qc.histogram_gc_content()
"""
pylab.hist(self.gc_list, bins=range(0, 100))
pylab.grid()
pylab.title("GC content distribution (per sequence)")
pylab.xlabel(r"Mean GC content (%)", fontsize=self.fontsize)
pylab.xlim([0,100])
def hist_GC(self, bins=50, alpha=0.5, hold=False, fontsize=12,
grid=True, xlabel="GC %", ylabel="#", label="",title=None):
"""Plot histogram GC content
:param int bins: binning for the histogram
:param float alpha: transparency of the histograms
:param bool hold:
:param int fontsize: fontsize of the x and y labels and title.
:param bool grid: add grid or not
:param str xlabel:
:param str ylabel:
:param str label: label of the histogram (for the legend)
: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_GC()
"""
mean_GC = np.mean(self.df.loc[:,'GC_content'])
# set title if needed
if title is None:
title = "GC %% \n Mean GC : %.2f" %(mean_GC)
# histogram GC percent
if hold is False:
pylab.clf()
pylab.hist(self.df.loc[:,'GC_content'], bins=bins,
alpha=alpha, label=label + ", mean : " + str(round(mean_GC, 2))
+ ", N : " + str(len(self)))
pylab.xlabel(xlabel, fontsize=fontsize)
pylab.ylabel(ylabel, fontsize=fontsize)
pylab.title(title, fontsize=fontsize)
if grid is True:
pylab.grid(True)
pylab.xlim([0, 100])
try: pylab.tight_layout()
except:pass
def plot_acgt_content(self, stacked=False):
"""Plot histogram of GC content
.. plot::
:include-source:
from sequana import sequana_data
from sequana import FastQC
filename = sequana_data("test.fastq", "testing")
qc = FastQC(filename)
qc.plot_acgt_content()
"""
df = self.get_actg_content()
if stacked is True:
df.plot.bar(stacked=True)
else:
df.plot()
pylab.grid(True)
pylab.xlabel("position (bp)", fontsize=self.fontsize)
pylab.ylabel("percent", fontsize=self.fontsize)
def hist_nb_passes(self, bins=None, alpha=0.5, hold=False, fontsize=12,
grid=True, xlabel="Number of ZMW passes", logy=True,
ylabel="#", label="", title="Number of ZMW passes"):
"""Plot histogram of number of reads per ZMW (number of passes)
:param float alpha: transparency of the histograms
:param bool hold:
:param int fontsize:
:param bool grid:
:param str xlabel:
:param str ylabel:
:param bool logy: use log scale on the y axis (default to True)
:param str label: label of the histogram (for the legend)
: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_nb_passes()
"""
max_nb_pass = self.df.nb_passes.max()
if bins is None:
k = range(1, max_nb_pass+1)
# histogram nb passes
if hold is False:
pylab.clf()
pylab.hist(self.df.nb_passes, bins=bins, alpha=alpha,
label=label, log=logy, width=1)
if len(k) < 5:
pylab.xticks(range(6), range(6))
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, bins=80, rwidth=0.8, **kwargs):
pylab.clf()
Y, X, _ = pylab.hist(self.data, bins=bins, rwidth=rwidth, **kwargs)
pylab.xlabel(self.xlabel, fontsize=self.fontsize)
pylab.ylabel(self.ylabel, fontsize=self.fontsize)
"""self.Y = Y
self.X = X
ax_twin = pylab.gca().twinx()
shift = (X[1] - X[0]) / 2
ax_twin.plot(X[0:-1]- shift, len(self.data) - pylab.cumsum(Y), "k")
ax_twin.set_ylim(bottom=0)
pylab.ylabel("CDF", fontsize=self.fontsize)
"""
pylab.grid(self.grid)
pylab.title(self.title)
try: pylab.tight_layout()
except:pass
def plot_GC_read_len(self, hold=False, fontsize=12, bins=[200, 60],
grid=True, xlabel="GC %", ylabel="#", cmap="BrBG"):
"""Plot GC content versus read length
:param bool hold:
:param int fontsize: for x and y labels and title
:param bins: a integer or tuple of 2 integers to specify
the binning of the x and y 2D histogram.
:param bool grid:
:param str xlabel:
:param str ylabel:
.. plot::
:include-source:
from sequana.pacbio import PacbioSubreads
from sequana import sequana_data
b = PacbioSubreads(sequana_data("test_pacbio_subreads.bam"))
b.plot_GC_read_len(bins=[10, 10])
"""
mean_len = np.mean(self.df.loc[:,'read_length'])
mean_GC = np.mean(self.df.loc[:,'GC_content'])
if hold is False:
pylab.clf()
data = self.df.loc[:,['read_length','GC_content']].dropna()
h = biokit.viz.hist2d.Hist2D(data)
res = h.plot(bins=bins, contour=False, norm='log', Nlevels=6, cmap=cmap)
pylab.xlabel("Read length", fontsize=fontsize)
pylab.ylabel("GC %", fontsize=fontsize)
pylab.title("GC %% vs length \n Mean length : %.2f , Mean GC : %.2f" %
(mean_len, mean_GC), fontsize=fontsize)
pylab.ylim([0, 100])
if grid is True:
pylab.grid(True)
