def venn(self, compa_list, direction="all", prefix=""):
"""
Plot a venn diagram comparing the list compa_list of dr gene lists.
compa_list is a list of comparison names from Deseq2 results
direction specifies either if up/down/all dr genes are considered
prefix is a string to be added as prefix to the outfile name.
compa_list can be a list of lists of comparisons to make.
ie [["WT", "KO1"],["WT", "KO2"]
"""
from sequana.viz.venn import plot_venn
# If compa_list is a list of lists of comparison
if all(isinstance(l, list) for l in compa_list):
fig, ax = pylab.subplots(6, 1, figsize=(6, 20))
ax = ax.flat
for i, c in enumerate(compa_list):
plot_venn(
[self.dr_gene_lists[x][direction] for x in c],
[compa_name for compa_name in c],
ax=ax[i],
)
# If compa is only a list of comparisons
else:
plot_venn(
[self.dr_gene_lists[x][direction] for x in compa_list],
[compa_name for compa_name in compa_list],
)
out_dir = os.path.join(self.out_dir, "vennDiagrams")
os.makedirs(out_dir, exist_ok=True)
outfile = os.path.join(out_dir, f"{prefix}vennDiagrams_{direction}.pdf")
pylab.savefig(outfile, bbox_inches="tight")
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 plot_rank_vs_idr_score(self, filename=None, savefig=False):
# rank versus IDR scores
f, axes = pylab.subplots(2, 1)
df = self.df
axes[0].plot(
range(len(df)),
df.sort_values(by='rep1_rank', ascending=False)['local_idr'], 'o')
axes[0].set_ylabel("log10 IDR for replicate 1")
axes[0].axvline(len(self.df) - self.N_significant_peaks,
color='b',
ls='--')
axes[1].plot(
range(len(df)),
df.sort_values(by='rep2_rank', ascending=False)['local_idr'], 'ro')
axes[1].set_ylabel("log10 IDR for replicate 2")
axes[1].axvline(len(self.df) - self.N_significant_peaks,
color='b',
ls='--')
if savefig:
pylab.savefig(filename)
def plot_all_skews(self, figsize=(10, 12), fontsize=16, alpha=0.5):
if self._window is None:
raise AttributeError("Please set a valid window to compute skew")
# create figure
# fig, axarr = pylab.subplots(10,1, sharex=True, figsize=figsize)
fig, axarr = pylab.subplots(9, 1, sharex=True, figsize=figsize)
main_title = "Window size = %d (%.0f %% of genome )\n\
GC content = %.0f %%, AT content = %.0f %%, ignored = %.0f %%" \
% (self._window, self._window*100/self.__len__(),
self.gc_content()*100, (1-self.gc_content())*100, self._ignored_nuc*100)
pylab.suptitle(main_title, fontsize=fontsize)
# GC skew
axarr[0].set_title("GC skew (blue) - Cumulative sum (red)")
axarr[0].plot(list(self._GC_skew_slide[0]), 'b-', alpha=alpha)
axarr[0].set_ylabel("(G -C) / (G + C)")
axarr[1].plot(list(np.cumsum(self._GC_skew_slide[0])),
'r-',
alpha=alpha)
axarr[1].set_ylabel("(G -C) / (G + C)")
# AT skew
axarr[2].set_title("AT skew (blue) - Cumulative sum (red)")
axarr[2].plot(list(self._AT_skew_slide[0]), 'b-', alpha=alpha)
axarr[2].set_ylabel("(A -T) / (A + T)")
axarr[3].plot(list(np.cumsum(self._AT_skew_slide[0])),
'r-',
alpha=alpha)
axarr[3].set_ylabel("(A -T) / (A + T)", rotation=0)
# Xn
axarr[4].set_title("Cumulative RY skew (Purine - Pyrimidine)")
axarr[4].plot(self._Xn, 'g-', alpha=alpha)
axarr[4].set_ylabel("(A + G) - (C + T)")
# Yn
axarr[5].set_title("Cumulative MK skew (Amino - Keto)")
axarr[5].plot(self._Yn, 'g-', alpha=alpha)
axarr[5].set_ylabel("(A + C) - (G + T)")
# Zn
axarr[6].set_title(
"Cumulative H-bond skew (Weak H-bond - Strong H-bond)")
axarr[6].plot(self._Zn, 'g-', alpha=alpha)
axarr[6].set_ylabel("(A + T) - (G + C)")
# GC content
axarr[7].set_title("GC content")
axarr[7].plot(list(self._GC_content_slide[0]), 'k-', alpha=alpha)
axarr[7].set_ylabel("GC")
# AT content
axarr[8].set_title("AT content")
axarr[8].plot(list(self._AT_content_slide[0]), 'k-', alpha=alpha)
axarr[8].set_ylabel("AT")
# # FFT
# axarr[9].set_title("FFT")
# axarr[9].plot(list(self._c_fft),'g-',alpha=alpha)
# axarr[9].set_ylabel("FFT")
fig.tight_layout()
fig.subplots_adjust(top=0.88)
shuffle(shuffle_col)
colors = [cmap(i) for i in shuffle_col]
pylab.plot(res_best["qLength"], res_best["score_norm"], "bo", alpha=0.5)
pylab.xlabel("Length of contig")
pylab.ylabel("Score blasr (normalised by length)")
pylab.title(title_plot)
if save_plot:
pylab.savefig(file_plot.replace(".png", "_scores.png"))
else:
pylab.show()
##### Plot by reference
ref_found = list(res_best["reference"].unique())
fig, axarr = pylab.subplots(2 * len(ref_found),
figsize=(15, 8 * len(ref_found)))
#fig.suptitle("Coverage by contigs (blasr)\n%s" % title_plot, fontsize=10)
for i in range(len(ref_found)):
# keep only contigs aligned on this reference
res_best_ref = res_best[res_best["reference"] == ref_found[i]]
len_genome = df_genome_len.loc[df_genome_len["name"] == ref_found[i],
"length"].values[0]
# plot coverage found by blasr, with score
ax = axarr[i * 2]
list_contigs = plot_contigs(res_best_ref, ax, colors, mode="score")
genome_not_covered = areas_not_covered(list_contigs, len_genome)
# add grey on not covered areas
for area in genome_not_covered:
ax.axvspan(area[0], area[1], alpha=0.1, color='k')
list_input = [name.split('\n')[0] for name in f.readlines()]
f.close()
# list of labels
f = open(labels_input, 'r')
list_labels = [name.split('\n')[0] for name in f.readlines()]
f.close()
################################ EXECUTE ##############################################################################################
#pylab.figure(figsize=(5, 5))
fig1, ax1 = pylab.subplots(1,1, figsize=(5, 5))
fig2, ax2 = pylab.subplots(1,1, figsize=(5, 5))
res_PS = []
for i in range(len(list_input)):
df = pd.read_csv(list_input[i])
##### Precision without unknown taxons
good_class_rank = df["good_classification_at_level"].sum()
tot_class_rank = df["good_classification_at_level"].sum() + df["wrong_classification_at_level"].sum()
tot = df["total_N_reads"].sum()
wrong_class_above = df["wrong_classification_above_level"].sum()
##### Precision without unknown taxons
# some reads are classified but we dont find any info : cannot ignore them
#colors = [cmap(i) for i in np.linspace(0,1,len(list_analysis))]
# positions of genome
gen_pos = [[i, i + step - 1] for i in range(0, len_genome, step)]
y_pos = list(np.linspace(0, 1, len(analysis_names) + 2))
if custom_colors:
y_col = [colors[i] for i in range(len(analysis_names))]
else:
y_col = [cmap(i) for i in np.linspace(0, 1, len(analysis_names))]
pylab.close('all')
# create figure
fig, axarr = pylab.subplots(len(gen_pos),
1,
figsize=(int(step / 20000),
int(len(gen_pos)) * 1.1))
for i in range(len(gen_pos)):
subplot_variant_position(df_result, i, gen_pos, axarr, analysis_names,
y_pos, y_col, be_repeats_concat)
# add grey at the end (no genome)
ax = axarr[-1]
ax.axvspan(len_genome, gen_pos[-1][1], alpha=0.5, color='k')
#fig.subplots_adjust(bottom=0.2)
#fig.tight_layout()
pylab.subplots_adjust(hspace=hspace_subplots)
pylab.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)
#pylab.legend(loc="lower left")
if file_fig != "show":
cmap = pylab.cm.get_cmap(colormap)
# shuffle colors : in case 2 adjacent contigs have the same color, user can plot again to see better
shuffle_col = list(np.linspace(0,1,res_best.shape[0]))
shuffle(shuffle_col)
colors = [cmap(i) for i in shuffle_col]
pylab.plot(res_best["qLength"], res_best["score_norm"],"bo",alpha=0.5)
pylab.xlabel("Length of contig")
pylab.ylabel("Score blasr (normalised by length)")
pylab.title(title_plot)
if save_plot:
pylab.savefig(file_plot.replace(".png","_scores.png"))
else:
pylab.show()
fig, axarr = pylab.subplots(2,figsize=figsize, sharex=True)
fig.suptitle("Coverage by contigs (blasr)\n%s" % title_plot, fontsize=10)
# plot coverage found by blasr, with score
ax = axarr[0]
list_contigs = plot_contigs(res_best, ax, mode="score")
genome_not_covered = areas_not_covered(list_contigs, len_genome)
# add grey on not covered areas
for area in genome_not_covered:
ax.axvspan(area[0],area[1], alpha=0.1, color='k')
ax.set_ylabel("Score blasr (normalised by length)")
# plot coverage found by blasr, with random y distribution (to see if there are overlaps)
ax = axarr[1]
list_contigs = plot_contigs(res_best, ax, mode="random")
for area in genome_not_covered:
ax.axvspan(area[0],area[1], alpha=0.05, color='k')
# plot score nor normalised
pylab.plot(res_blasr["qLength"], res_blasr["score"], "bo", alpha=0.5)
pylab.xlabel("Length of contig")
pylab.ylabel("Score blasr (not normalised)")
pylab.title(title_plot)
pylab.show()
# plot score normalised by lenght
pylab.plot(res_blasr["qLength"], res_blasr["score_norm"], "bo", alpha=0.5)
pylab.xlabel("Length of contig")
pylab.ylabel("Score blasr (normalised by length)")
pylab.title(title_plot)
pylab.show()
fig, axarr = pylab.subplots(2, figsize=(15, 8), sharex=True)
fig.suptitle("Coverage by contigs (blasr)\n%s" % title_plot, fontsize=10)
# plot coverage found by blasr, with score
ax = axarr[0]
for i in range(res_blasr.shape[0]):
res_to_plot = res_blasr.iloc[i, :]
contig = res_to_plot["qName"]
start = int(res_to_plot["tStart"])
end = int(res_to_plot["tEnd"])
score = float(res_to_plot["score_norm"])
ax.plot([start, end], [score] * 2,
ls='-',
lw=5,
color=colors[i],
solid_capstyle="butt")
ax.set_ylabel("Score blasr (normalised by length)")
################################ IMPORT DATA ##############################################################################################
list_files = read_fof(fof_BAM)
description = pd.read_csv(file_description)
list_BAM = []
labels = []
for f in list_files:
short_name_bam = f.split("/")[-1]
labels.append(description[description["Filename"] == short_name_bam]
["polymerase"].values[0])
list_BAM.append(pacbio.BAMPacbio(f))
# plot read length
fig, ax = pylab.subplots(1, 1, figsize=figsize_read_len)
for i in range(len(list_BAM)):
bam = list_BAM[i]
bam.hist_len(hold=True, grid=False, label=labels[i], title="")
ax.legend()
fig.tight_layout()
if save_plots:
pylab.savefig(filename_output.replace(".", "_read_len."), dpi=182)
pylab.clf()
else:
pylab.show()
# plot GC %
fig, ax = pylab.subplots(1, 1, figsize=figsize_GC)
for i in range(len(list_BAM)):
bam = list_BAM[i]
def plot_all_skews(self,figsize=(10, 12), fontsize=16, alpha=0.5):
if self._window is None:
raise AttributeError("Please set a valid window to compute skew")
# create figure
# fig, axarr = pylab.subplots(10,1, sharex=True, figsize=figsize)
fig, axarr = pylab.subplots(9,1, sharex=True, figsize=figsize)
main_title = "Window size = %d (%.0f %% of genome )\n\
GC content = %.0f %%, AT content = %.0f %%, ignored = %.0f %%" \
% (self._window, self._window*100/self.__len__(),
self.gc_content()*100, (1-self.gc_content())*100, self._ignored_nuc*100)
pylab.suptitle(main_title, fontsize=fontsize)
# GC skew
axarr[0].set_title("GC skew (blue) - Cumulative sum (red)")
axarr[0].plot(list(self._GC_skew_slide[0]),'b-',alpha=alpha)
axarr[0].set_ylabel("(G -C) / (G + C)")
axarr[1].plot(list(np.cumsum(self._GC_skew_slide[0])),'r-',alpha=alpha)
axarr[1].set_ylabel("(G -C) / (G + C)")
# AT skew
axarr[2].set_title("AT skew (blue) - Cumulative sum (red)")
axarr[2].plot(list(self._AT_skew_slide[0]),'b-',alpha=alpha)
axarr[2].set_ylabel("(A -T) / (A + T)")
axarr[3].plot(list(np.cumsum(self._AT_skew_slide[0])),'r-',alpha=alpha)
axarr[3].set_ylabel("(A -T) / (A + T)", rotation=0)
# Xn
axarr[4].set_title("Cumulative RY skew (Purine - Pyrimidine)")
axarr[4].plot(self._Xn,'g-',alpha=alpha)
axarr[4].set_ylabel("(A + G) - (C + T)")
# Yn
axarr[5].set_title("Cumulative MK skew (Amino - Keto)")
axarr[5].plot(self._Yn,'g-',alpha=alpha)
axarr[5].set_ylabel("(A + C) - (G + T)")
# Zn
axarr[6].set_title("Cumulative H-bond skew (Weak H-bond - Strong H-bond)")
axarr[6].plot(self._Zn,'g-',alpha=alpha)
axarr[6].set_ylabel("(A + T) - (G + C)")
# GC content
axarr[7].set_title("GC content")
axarr[7].plot(list(self._GC_content_slide[0]),'k-',alpha=alpha)
axarr[7].set_ylabel("GC")
# AT content
axarr[8].set_title("AT content")
axarr[8].plot(list(self._AT_content_slide[0]),'k-',alpha=alpha)
axarr[8].set_ylabel("AT")
# # FFT
# axarr[9].set_title("FFT")
# axarr[9].plot(list(self._c_fft),'g-',alpha=alpha)
# axarr[9].set_ylabel("FFT")
fig.tight_layout()
fig.subplots_adjust(top=0.88)