def resample(args):
"""
%prog resample yellow-catfish-resample.txt medicago-resample.txt
Plot ALLMAPS performance across resampled real data.
"""
p = OptionParser(resample.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="8x4", dpi=300)
if len(args) != 2:
sys.exit(not p.print_help())
dataA, dataB = args
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
A = fig.add_axes([.1, .18, .32, .64])
B = fig.add_axes([.6, .18, .32, .64])
dataA = import_data(dataA)
dataB = import_data(dataB)
xlabel = "Fraction of markers"
ylabels = ("Anchor rate", "Runtime (m)")
legend = ("anchor rate", "runtime")
subplot_twinx(A, dataA, xlabel, ylabels,
title="Yellow catfish", legend=legend)
subplot_twinx(B, dataB, xlabel, ylabels,
title="Medicago", legend=legend)
labels = ((.04, .92, "A"), (.54, .92, "B"))
panel_labels(root, labels)
normalize_axes(root)
image_name = "resample." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def allelefreq(args):
"""
%prog allelefreq HD,DM1,SCA1,SCA17
Plot the allele frequencies of some STRs.
"""
p = OptionParser(allelefreq.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="10x10")
if len(args) != 1:
sys.exit(not p.print_help())
loci, = args
fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(ncols=2,
nrows=2,
figsize=(iopts.w, iopts.h))
plt.tight_layout(pad=4)
treds, df = read_treds()
df = df.set_index(["abbreviation"])
for ax, locus in zip((ax1, ax2, ax3, ax4), loci.split(",")):
plot_allelefreq(ax, df, locus)
root = fig.add_axes([0, 0, 1, 1])
pad = .03
panel_labels(root, ((pad / 2, 1 - pad, "A"), (1 / 2., 1 - pad, "B"),
(pad / 2, 1 / 2., "C"), (1 / 2., 1 / 2., "D")))
normalize_axes(root)
image_name = "allelefreq." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def composite_correlation(df, size=(12, 8)):
""" Plot composite correlation figure
"""
fig = plt.figure(1, size)
ax1 = plt.subplot2grid((2, 2), (0, 0))
ax2 = plt.subplot2grid((2, 2), (0, 1))
ax3 = plt.subplot2grid((2, 2), (1, 0))
ax4 = plt.subplot2grid((2, 2), (1, 1))
chemistry = ["V1", "V2", "V2.5", float("nan")]
colors = sns.color_palette("Set2", 8)
color_map = dict(zip(chemistry, colors))
age_label = "Chronological age (yr)"
ax1.scatter(df["hli_calc_age_sample_taken"],
df["teloLength"],
s=10,
marker='.',
color=df["Chemistry"].map(color_map))
ax1.set_ylim(0, 15)
ax1.set_ylabel("Telomere length (Kb)")
ax2.scatter(df["hli_calc_age_sample_taken"],
df["ccn.chrX"],
s=10,
marker='.',
color=df["Chemistry"].map(color_map))
ax2.set_ylim(1.8, 2.1)
ax2.set_ylabel("ChrX copy number")
ax4.scatter(df["hli_calc_age_sample_taken"],
df["ccn.chrY"],
s=10,
marker='.',
color=df["Chemistry"].map(color_map))
ax4.set_ylim(0.8, 1.1)
ax4.set_ylabel("ChrY copy number")
ax3.scatter(df["hli_calc_age_sample_taken"],
df["TRA.PPM"],
s=10,
marker='.',
color=df["Chemistry"].map(color_map))
ax3.set_ylim(0, 250)
ax3.set_ylabel("$TCR-\\alpha$ deletions (count per million reads)")
from matplotlib.lines import Line2D
legend_elements = [Line2D([0], [0], marker='.', color='w', label=chem,
markerfacecolor=color, markersize=16) \
for (chem, color) in zip(chemistry, colors)[:3]]
for ax in (ax1, ax2, ax3, ax4):
ax.set_xlabel(age_label)
ax.legend(handles=legend_elements, loc="upper right")
plt.tight_layout()
root = fig.add_axes((0, 0, 1, 1))
labels = ((.02, .98, "A"), (.52, .98, "B"), (.02, .5, "C"), (.52, .5, "D"))
panel_labels(root, labels)
root.set_xlim(0, 1)
root.set_ylim(0, 1)
root.set_axis_off()
def composite_qc(df_orig, size=(16, 12)):
""" Plot composite QC figures
"""
df = df_orig.rename(columns={"hli_calc_age_sample_taken": "Age",
"hli_calc_gender": "Gender",
"eth7_max": "Ethnicity",
"MeanCoverage": "Mean coverage",
"Chemistry": "Sequencing chemistry",
"Release Client": "Cohort",
})
fig = plt.figure(1, size)
ax1 = plt.subplot2grid((2, 7), (0, 0), rowspan=1, colspan=2)
ax2 = plt.subplot2grid((2, 7), (0, 2), rowspan=1, colspan=2)
ax3 = plt.subplot2grid((2, 7), (0, 4), rowspan=1, colspan=3)
ax4 = plt.subplot2grid((2, 7), (1, 0), rowspan=1, colspan=2)
ax5 = plt.subplot2grid((2, 7), (1, 2), rowspan=1, colspan=2)
ax6 = plt.subplot2grid((2, 7), (1, 4), rowspan=1, colspan=3)
sns.distplot(df["Age"].dropna(), kde=False, ax=ax1)
sns.countplot(x="Gender", data=df, ax=ax2)
sns.countplot(x="Ethnicity", data=df, ax=ax3,
order = df['Ethnicity'].value_counts().index)
sns.distplot(df["Mean coverage"].dropna(), kde=False, ax=ax4)
ax4.set_xlim(0, 100)
sns.countplot(x="Sequencing chemistry", data=df, ax=ax5)
sns.countplot(x="Cohort", data=df, ax=ax6,
order = df['Cohort'].value_counts().index)
# Anonymize the cohorts
cohorts = ax6.get_xticklabels()
newCohorts = []
for i, c in enumerate(cohorts):
if c.get_text() == "Spector":
c = "TwinsUK"
elif c.get_text() != "Health Nucleus":
c = "C{}".format(i + 1)
newCohorts.append(c)
ax6.set_xticklabels(newCohorts)
for ax in (ax6,):
ax.set_xticklabels(ax.get_xticklabels(), ha="right", rotation=30)
for ax in (ax1, ax2, ax3, ax4, ax5, ax6):
ax.set_title(ax.get_xlabel())
ax.set_xlabel("")
plt.tight_layout()
root = fig.add_axes((0, 0, 1, 1))
labels = ((.02, .96, "A"),
(.3, .96, "B"),
(.6, .96, "C"),
(.02, .52, "D"),
(.3, .52, "E"),
(.6, .52, "F"))
panel_labels(root, labels)
root.set_xlim(0, 1)
root.set_ylim(0, 1)
root.set_axis_off()
def pomegranate(args):
"""
%prog cotton seqids karyotype.layout mcscan.out all.bed synteny.layout
Build a figure that calls graphics.karyotype to illustrate the high ploidy
of WGD history of pineapple genome. The script calls both graphics.karyotype
and graphic.synteny.
"""
p = OptionParser(pomegranate.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="9x7")
if len(args) != 5:
sys.exit(not p.print_help())
seqidsfile, klayout, datafile, bedfile, slayout = args
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
Karyotype(fig, root, seqidsfile, klayout)
Synteny(fig, root, datafile, bedfile, slayout)
# legend showing the orientation of the genes
draw_gene_legend(root, 0.42, 0.52, 0.48)
labels = ((0.04, 0.96, "A"), (0.04, 0.52, "B"))
panel_labels(root, labels)
root.set_xlim(0, 1)
root.set_ylim(0, 1)
root.set_axis_off()
pf = "pomegranate-karyotype"
image_name = pf + "." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def pomegranate(args):
"""
%prog cotton seqids karyotype.layout mcscan.out all.bed synteny.layout
Build a figure that calls graphics.karyotype to illustrate the high ploidy
of WGD history of pineapple genome. The script calls both graphics.karyotype
and graphic.synteny.
"""
p = OptionParser(pomegranate.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="9x7")
if len(args) != 5:
sys.exit(not p.print_help())
seqidsfile, klayout, datafile, bedfile, slayout = args
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
Karyotype(fig, root, seqidsfile, klayout)
Synteny(fig, root, datafile, bedfile, slayout)
# legend showing the orientation of the genes
draw_gene_legend(root, .42, .52, .48)
labels = ((.04, .96, 'A'), (.04, .52, 'B'))
panel_labels(root, labels)
root.set_xlim(0, 1)
root.set_ylim(0, 1)
root.set_axis_off()
pf = "pomegranate-karyotype"
image_name = pf + "." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def resample(args):
"""
%prog resample yellow-catfish-resample.txt medicago-resample.txt
Plot ALLMAPS performance across resampled real data.
"""
p = OptionParser(resample.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="8x4", dpi=300)
if len(args) != 2:
sys.exit(not p.print_help())
dataA, dataB = args
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
A = fig.add_axes([0.1, 0.18, 0.32, 0.64])
B = fig.add_axes([0.6, 0.18, 0.32, 0.64])
dataA = import_data(dataA)
dataB = import_data(dataB)
xlabel = "Fraction of markers"
ylabels = ("Anchor rate", "Runtime (m)")
legend = ("anchor rate", "runtime")
subplot_twinx(A, dataA, xlabel, ylabels, title="Yellow catfish", legend=legend)
subplot_twinx(B, dataB, xlabel, ylabels, title="Medicago", legend=legend)
labels = ((0.04, 0.92, "A"), (0.54, 0.92, "B"))
panel_labels(root, labels)
normalize_axes(root)
image_name = "resample." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def compare2(args):
"""
%prog compare2
Compare performances of various variant callers on simulated STR datasets.
"""
p = OptionParser(compare2.__doc__)
p.add_option('--maxinsert',
default=300,
type="int",
help="Maximum number of repeats")
add_simulate_options(p)
opts, args, iopts = p.set_image_options(args, figsize="10x5")
if len(args) != 0:
sys.exit(not p.print_help())
depth = opts.depth
readlen = opts.readlen
distance = opts.distance
max_insert = opts.maxinsert
fig, (ax1, ax2) = plt.subplots(ncols=2,
nrows=1,
figsize=(iopts.w, iopts.h))
plt.tight_layout(pad=2)
# ax1: lobSTR vs TREDPARSE with haploid model
lobstr_results = parse_results("lobstr_results_homo.txt")
tredparse_results = parse_results("tredparse_results_homo.txt")
title = SIMULATED_HAPLOID + \
r" ($D=%s\times, L=%dbp, V=%dbp$)" % (depth, readlen, distance)
plot_compare(ax1,
title,
tredparse_results,
lobstr_results,
max_insert=max_insert)
# ax2: lobSTR vs TREDPARSE with diploid model
lobstr_results = parse_results("lobstr_results_het.txt", exclude=20)
tredparse_results = parse_results("tredparse_results_het.txt", exclude=20)
title = SIMULATED_DIPLOID + \
r" ($D=%s\times, L=%dbp, V=%dbp$)" % (depth, readlen, distance)
plot_compare(ax2,
title,
tredparse_results,
lobstr_results,
max_insert=max_insert)
for ax in (ax1, ax2):
ax.set_xlim(0, max_insert)
ax.set_ylim(0, max_insert)
root = fig.add_axes([0, 0, 1, 1])
pad = .03
panel_labels(root, ((pad / 2, 1 - pad, "A"), (1 / 2., 1 - pad, "B")))
normalize_axes(root)
image_name = "tredparse." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def compare(args):
"""
%prog compare Evaluation.csv
Compare performances of various variant callers on simulated STR datasets.
"""
p = OptionParser(compare.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="10x10")
if len(args) != 1:
sys.exit(not p.print_help())
datafile, = args
pf = datafile.rsplit(".", 1)[0]
fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(ncols=2,
nrows=2,
figsize=(iopts.w, iopts.h))
plt.tight_layout(pad=3)
bbox = {'facecolor': 'tomato', 'alpha': .2, 'ec': 'w'}
pad = 2
# Read benchmark data
df = pd.read_csv("Evaluation.csv")
truth = df["Truth"]
axes = (ax1, ax2, ax3, ax4)
progs = ("Manta", "Isaac", "GATK", "lobSTR")
markers = ("bx-", "yo-", "md-", "c+-")
for ax, prog, marker in zip(axes, progs, markers):
ax.plot(truth, df[prog], marker)
ax.plot(truth, truth, 'k--') # to show diagonal
ax.axhline(infected_thr, color='tomato')
ax.text(max(truth) - pad,
infected_thr + pad,
'Risk threshold',
bbox=bbox,
ha="right")
ax.axhline(ref_thr, color='tomato')
ax.text(max(truth) - pad,
ref_thr - pad,
'Reference repeat count',
bbox=bbox,
ha="right",
va="top")
ax.set_title(SIMULATED_HAPLOID)
ax.set_xlabel(r'Num of CAG repeats inserted ($\mathit{h}$)')
ax.set_ylabel('Num of CAG repeats called')
ax.legend([prog, 'Truth'], loc='best')
root = fig.add_axes([0, 0, 1, 1])
pad = .03
panel_labels(root, ((pad / 2, 1 - pad, "A"), (1 / 2., 1 - pad, "B"),
(pad / 2, 1 / 2., "C"), (1 / 2., 1 / 2., "D")))
normalize_axes(root)
image_name = pf + "." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def composite(df, sameGenderMZ, sameGenderDZ, size=(16, 24)):
"""Embed both absdiff figures and heritability figures.
"""
fig = plt.figure(1, size)
ax1a = plt.subplot2grid((6, 4), (0, 0), rowspan=2, colspan=1)
ax2a = plt.subplot2grid((6, 4), (0, 1), rowspan=2, colspan=1)
ax3a = plt.subplot2grid((6, 4), (0, 2), rowspan=2, colspan=1)
ax4a = plt.subplot2grid((6, 4), (0, 3), rowspan=2, colspan=1)
ax1b = plt.subplot2grid((6, 4), (2, 0), rowspan=2, colspan=2)
ax2b = plt.subplot2grid((6, 4), (2, 2), rowspan=2, colspan=2)
ax3b = plt.subplot2grid((6, 4), (4, 0), rowspan=2, colspan=2)
ax4b = plt.subplot2grid((6, 4), (4, 2), rowspan=2, colspan=2)
# Telomeres
telomeres = extract_trait(df, "Sample name", "telomeres.Length")
mzTelomeres = extract_twin_values(sameGenderMZ, telomeres)
dzTelomeres = extract_twin_values(sameGenderDZ, telomeres)
plot_paired_values(ax1b, mzTelomeres, dzTelomeres, label="Telomere length")
plot_abs_diff(ax1a, mzTelomeres, dzTelomeres, label="Telomere length")
# CCNX
CCNX = extract_trait(df, "Sample name", "ccn.chrX")
mzCCNX = extract_twin_values(sameGenderMZ, CCNX, gender="Female")
dzCCNX = extract_twin_values(sameGenderDZ, CCNX, gender="Female")
dzCCNX = filter_low_values(dzCCNX, 1.75)
plot_paired_values(ax2b, mzCCNX, dzCCNX, gender="Female only", label="ChrX copy number")
plot_abs_diff(ax2a, mzCCNX, dzCCNX, label="ChrX copy number")
# CCNY
CCNY = extract_trait(df, "Sample name", "ccn.chrY")
mzCCNY = extract_twin_values(sameGenderMZ, CCNY, gender="Male")
dzCCNY = extract_twin_values(sameGenderDZ, CCNY, gender="Male")
dzCCNY = filter_low_values(dzCCNY, .75)
plot_paired_values(ax3b, mzCCNY, dzCCNY, gender="Male only", label="ChrY copy number")
plot_abs_diff(ax3a, mzCCNY, dzCCNY, label="ChrY copy number")
# CCNY
TRA = extract_trait(df, "Sample name", "TRA.PPM")
mzTRA = extract_twin_values(sameGenderMZ, TRA)
dzTRA = extract_twin_values(sameGenderDZ, TRA)
plot_paired_values(ax4b, mzTRA, dzTRA, label="TCR-$\\alpha$ deletions")
plot_abs_diff(ax4a, mzTRA, dzTRA, label="TCR-$\\alpha$ deletions")
plt.tight_layout()
root = fig.add_axes((0, 0, 1, 1))
# ABCD absdiff, EFGH heritability
labels = ((.03, .99, 'A'), (.27, .99, 'B'), (.53, .99, 'C'), (.77, .99, 'D'),
(.03, .67, 'E'), (.53, .67, 'F'),
(.03, .34, 'G'), (.53, .34, 'H'))
panel_labels(root, labels)
root.set_xlim(0, 1)
root.set_ylim(0, 1)
root.set_axis_off()
def venn(args):
"""
%prog venn *.benchmark
Display benchmark results as Venn diagram.
"""
from matplotlib_venn import venn2
p = OptionParser(venn.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="9x9")
if len(args) < 1:
sys.exit(not p.print_help())
bcs = args
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
pad = .02
ystart = 1
ywidth = 1. / len(bcs)
tags = ("Bowers", "YGOB", "Schnable")
for bc, tag in zip(bcs, tags):
fp = open(bc)
data = []
for row in fp:
prog, pcounts, tcounts, shared = row.split()
pcounts = int(pcounts)
tcounts = int(tcounts)
shared = int(shared)
data.append((prog, pcounts, tcounts, shared))
xstart = 0
xwidth = 1. / len(data)
for prog, pcounts, tcounts, shared in data:
a, b, c = pcounts - shared, tcounts - shared, shared
ax = fig.add_axes([xstart + pad, ystart - ywidth + pad,
xwidth - 2 * pad, ywidth - 2 * pad])
venn2(subsets=(a, b, c), set_labels=(prog, tag), ax=ax)
message = "Sn={0} Pu={1}".\
format(percentage(shared, tcounts, precision=0, mode=-1),
percentage(shared, pcounts, precision=0, mode=-1))
print >> sys.stderr, message
ax.text(.5, .92, latex(message), ha="center", va="center",
transform=ax.transAxes, color='b')
ax.set_axis_off()
xstart += xwidth
ystart -= ywidth
panel_labels(root, ((.04, .96, "A"), (.04, .96 - ywidth, "B"),
(.04, .96 - 2 * ywidth, "C")))
panel_labels(root, ((.5, .98, "A. thaliana duplicates"),
(.5, .98 - ywidth, "14 Yeast genomes"),
(.5, .98 - 2 * ywidth, "4 Grass genomes")))
normalize_axes(root)
savefig("venn.pdf", dpi=opts.dpi)
def composite_correlation(df, size=(12, 8)):
""" Plot composite correlation figure
"""
fig = plt.figure(1, size)
ax1 = plt.subplot2grid((2, 2), (0, 0))
ax2 = plt.subplot2grid((2, 2), (0, 1))
ax3 = plt.subplot2grid((2, 2), (1, 0))
ax4 = plt.subplot2grid((2, 2), (1, 1))
chemistry = ["V1", "V2", "V2.5", float("nan")]
colors = sns.color_palette("Set2", 8)
color_map = dict(zip(chemistry, colors))
age_label = "Chronological age (yr)"
ax1.scatter(df["hli_calc_age_sample_taken"], df["teloLength"],
s=10, marker='.',
color=df["Chemistry"].map(color_map))
ax1.set_ylim(0, 15)
ax1.set_ylabel("Telomere length (Kb)")
ax2.scatter(df["hli_calc_age_sample_taken"], df["ccn.chrX"],
s=10, marker='.',
color=df["Chemistry"].map(color_map))
ax2.set_ylim(1.8, 2.1)
ax2.set_ylabel("ChrX copy number")
ax4.scatter(df["hli_calc_age_sample_taken"], df["ccn.chrY"],
s=10, marker='.',
color=df["Chemistry"].map(color_map))
ax4.set_ylim(0.8, 1.1)
ax4.set_ylabel("ChrY copy number")
ax3.scatter(df["hli_calc_age_sample_taken"], df["TRA.PPM"],
s=10, marker='.',
color=df["Chemistry"].map(color_map))
ax3.set_ylim(0, 250)
ax3.set_ylabel("$TCR-\\alpha$ deletions (count per million reads)")
from matplotlib.lines import Line2D
legend_elements = [Line2D([0], [0], marker='.', color='w', label=chem,
markerfacecolor=color, markersize=16) \
for (chem, color) in zip(chemistry, colors)[:3]]
for ax in (ax1, ax2, ax3, ax4):
ax.set_xlabel(age_label)
ax.legend(handles=legend_elements, loc="upper right")
plt.tight_layout()
root = fig.add_axes((0, 0, 1, 1))
labels = ((.02, .98, "A"),
(.52, .98, "B"),
(.02, .5, "C"),
(.52, .5, "D"))
panel_labels(root, labels)
root.set_xlim(0, 1)
root.set_ylim(0, 1)
root.set_axis_off()
def venn(args):
"""
%prog venn *.benchmark
Display benchmark results as Venn diagram.
"""
from matplotlib_venn import venn2
p = OptionParser(venn.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="9x9")
if len(args) < 1:
sys.exit(not p.print_help())
bcs = args
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
pad = .02
ystart = 1
ywidth = 1. / len(bcs)
tags = ("Bowers", "YGOB", "Schnable")
for bc, tag in zip(bcs, tags):
fp = open(bc)
data = []
for row in fp:
prog, pcounts, tcounts, shared = row.split()
pcounts = int(pcounts)
tcounts = int(tcounts)
shared = int(shared)
data.append((prog, pcounts, tcounts, shared))
xstart = 0
xwidth = 1. / len(data)
for prog, pcounts, tcounts, shared in data:
a, b, c = pcounts - shared, tcounts - shared, shared
ax = fig.add_axes([xstart + pad, ystart - ywidth + pad,
xwidth - 2 * pad, ywidth - 2 * pad])
venn2(subsets=(a, b, c), set_labels=(prog, tag), ax=ax)
message = "Sn={0} Pu={1}".\
format(percentage(shared, tcounts, precision=0, mode=-1),
percentage(shared, pcounts, precision=0, mode=-1))
print(message, file=sys.stderr)
ax.text(.5, .92, latex(message), ha="center", va="center",
transform=ax.transAxes, color='b')
ax.set_axis_off()
xstart += xwidth
ystart -= ywidth
panel_labels(root, ((.04, .96, "A"), (.04, .96 - ywidth, "B"),
(.04, .96 - 2 * ywidth, "C")))
panel_labels(root, ((.5, .98, "A. thaliana duplicates"),
(.5, .98 - ywidth, "14 Yeast genomes"),
(.5, .98 - 2 * ywidth, "4 Grass genomes")))
normalize_axes(root)
savefig("venn.pdf", dpi=opts.dpi)
def ploidy(args):
"""
%prog ploidy seqids karyotype.layout mcscan.out all.bed synteny.layout
Build a figure that calls graphics.karyotype to illustrate the high ploidy
of WGD history of pineapple genome. The script calls both graphics.karyotype
and graphic.synteny.
"""
p = OptionParser(ploidy.__doc__)
p.add_option("--switch", help="Rename the seqid with two-column file")
opts, args, iopts = p.set_image_options(args, figsize="9x7")
if len(args) != 5:
sys.exit(not p.print_help())
seqidsfile, klayout, datafile, bedfile, slayout = args
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
Karyotype(fig, root, seqidsfile, klayout)
Synteny(fig, root, datafile, bedfile, slayout, switch=opts.switch)
# legend showing the orientation of the genes
draw_gene_legend(root, 0.27, 0.37, 0.52)
# annotate the WGD events
fc = "lightslategrey"
x = 0.09
radius = 0.012
TextCircle(root, x, 0.825, r"$\tau$", radius=radius, fc=fc)
TextCircle(root, x, 0.8, r"$\sigma$", radius=radius, fc=fc)
TextCircle(root, x, 0.72, r"$\rho$", radius=radius, fc=fc)
for ypos in (0.825, 0.8, 0.72):
root.text(0.12, ypos, r"$\times2$", color=fc, ha="center", va="center")
root.plot([x, x], [0.85, 0.775], ":", color=fc, lw=2)
root.plot([x, x], [0.75, 0.675], ":", color=fc, lw=2)
labels = ((0.04, 0.96, "A"), (0.04, 0.54, "B"))
panel_labels(root, labels)
root.set_xlim(0, 1)
root.set_ylim(0, 1)
root.set_axis_off()
pf = "pineapple-karyotype"
image_name = pf + "." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def ploidy(args):
"""
%prog cotton seqids karyotype.layout mcscan.out all.bed synteny.layout
Build a figure that calls graphics.karyotype to illustrate the high ploidy
of WGD history of pineapple genome. The script calls both graphics.karyotype
and graphic.synteny.
"""
p = OptionParser(ploidy.__doc__)
p.add_option("--switch", help="Rename the seqid with two-column file")
opts, args, iopts = p.set_image_options(args, figsize="9x7")
if len(args) != 5:
sys.exit(not p.print_help())
seqidsfile, klayout, datafile, bedfile, slayout = args
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
Karyotype(fig, root, seqidsfile, klayout)
Synteny(fig, root, datafile, bedfile, slayout, switch=opts.switch)
# legend showing the orientation of the genes
draw_gene_legend(root, .27, .37, .52)
# annotate the WGD events
fc = 'lightslategrey'
x = .09
radius = .012
TextCircle(root, x, .825, r'$\tau$', radius=radius, fc=fc)
TextCircle(root, x, .8, r'$\sigma$', radius=radius, fc=fc)
TextCircle(root, x, .72, r'$\rho$', radius=radius, fc=fc)
for ypos in (.825, .8, .72):
root.text(.12, ypos, r"$\times2$", color=fc, ha="center", va="center")
root.plot([x, x], [.85, .775], ":", color=fc, lw=2)
root.plot([x, x], [.75, .675], ":", color=fc, lw=2)
labels = ((.04, .96, 'A'), (.04, .54, 'B'))
panel_labels(root, labels)
root.set_xlim(0, 1)
root.set_ylim(0, 1)
root.set_axis_off()
pf = "pineapple-karyotype"
image_name = pf + "." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def simulation(args):
"""
%prog simulation inversion.txt translocation.txt maps.txt multimaps.txt
Plot ALLMAPS accuracy across a range of simulated datasets.
"""
p = OptionParser(simulation.__doc__)
opts, args, iopts = p.set_image_options(args, dpi=300)
if len(args) != 4:
sys.exit(not p.print_help())
dataA, dataB, dataC, dataD = args
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
A = fig.add_axes([0.12, 0.62, 0.35, 0.35])
B = fig.add_axes([0.62, 0.62, 0.35, 0.35])
C = fig.add_axes([0.12, 0.12, 0.35, 0.35])
D = fig.add_axes([0.62, 0.12, 0.35, 0.35])
dataA = import_data(dataA)
dataB = import_data(dataB)
dataC = import_data(dataC)
dataD = import_data(dataD)
subplot(A, dataA, "Inversion error rate", "Accuracy", xlim=0.5)
subplot(
B,
dataB,
"Translocation error rate",
"Accuracy",
xlim=0.5,
legend=("intra-chromosomal", "inter-chromosomal", "75\% intra + 25\% inter"),
)
subplot(C, dataC, "Number of input maps", "Accuracy", xcast=int)
subplot(D, dataD, "Number of input maps", "Accuracy", xcast=int)
labels = (
(0.03, 0.97, "A"),
(0.53, 0.97, "B"),
(0.03, 0.47, "C"),
(0.53, 0.47, "D"),
)
panel_labels(root, labels)
normalize_axes(root)
image_name = "simulation." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def synteny(args):
"""
%prog synteny vplanifoliaA_blocks.bed vplanifoliaA.sizes \
b1.blocks all.bed b1.layout
Create a composite figure with (A) wgd and (B) microsynteny.
"""
from jcvi.graphics.chromosome import draw_chromosomes
p = OptionParser(synteny.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="12x12")
(bedfile, sizesfile, blocksfile, allbedfile, blockslayout) = args
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
ax1 = fig.add_axes([0, 0.5, 1, 0.5])
ax2 = fig.add_axes([0.02, 0, 0.98, 0.5])
# Panel A
title = r"Genome duplication $\alpha^{O}$ event in $\textit{Vanilla}$"
draw_chromosomes(
ax1,
bedfile,
sizes=sizesfile,
iopts=iopts,
mergedist=200000,
winsize=50000,
imagemap=False,
gauge=True,
legend=False,
title=title,
)
# Panel B
draw_ploidy(fig, ax2, blocksfile, allbedfile, blockslayout)
normalize_axes([root, ax1, ax2])
labels = ((0.05, 0.95, "A"), (0.05, 0.5, "B"))
panel_labels(root, labels)
image_name = "synteny.pdf"
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def likelihood3(args):
"""
%prog likelihood2 200_20.json 200_100.json
Plot the likelihood surface and marginal distributions for two settings.
"""
from matplotlib import gridspec
p = OptionParser(likelihood3.__doc__)
opts, args, iopts = p.set_image_options(args,
figsize="10x10",
style="white",
cmap="coolwarm")
if len(args) != 2:
sys.exit(not p.print_help())
jsonfile1, jsonfile2 = args
fig = plt.figure(figsize=(iopts.w, iopts.h))
gs = gridspec.GridSpec(9, 2)
ax1 = fig.add_subplot(gs[:4, 0])
ax2 = fig.add_subplot(gs[:2, 1])
ax3 = fig.add_subplot(gs[2:4, 1])
ax4 = fig.add_subplot(gs[5:, 0])
ax5 = fig.add_subplot(gs[5:7, 1])
ax6 = fig.add_subplot(gs[7:, 1])
plt.tight_layout(pad=2)
plot_panel(jsonfile1, ax1, ax2, ax3, opts.cmap)
plot_panel(jsonfile2, ax4, ax5, ax6, opts.cmap)
root = fig.add_axes([0, 0, 1, 1])
pad = .02
panel_labels(root, ((pad, 1 - pad, "A"), (pad, 4. / 9, "B")))
normalize_axes(root)
image_name = "likelihood3." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def simulation(args):
"""
%prog simulation inversion.txt translocation.txt maps.txt multimaps.txt
Plot ALLMAPS accuracy across a range of simulated datasets.
"""
p = OptionParser(simulation.__doc__)
opts, args, iopts = p.set_image_options(args, dpi=300)
if len(args) != 4:
sys.exit(not p.print_help())
dataA, dataB, dataC, dataD = args
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
A = fig.add_axes([.12, .62, .35, .35])
B = fig.add_axes([.62, .62, .35, .35])
C = fig.add_axes([.12, .12, .35, .35])
D = fig.add_axes([.62, .12, .35, .35])
dataA = import_data(dataA)
dataB = import_data(dataB)
dataC = import_data(dataC)
dataD = import_data(dataD)
subplot(A, dataA, "Inversion error rate", "Accuracy", xlim=.5)
subplot(B, dataB, "Translocation error rate", "Accuracy", xlim=.5,
legend=("intra-chromosomal", "inter-chromosomal",
"75\% intra + 25\% inter"))
subplot(C, dataC, "Number of input maps", "Accuracy", xcast=int)
subplot(D, dataD, "Number of input maps", "Accuracy", xcast=int)
labels = ((.03, .97, "A"), (.53, .97, "B"),
(.03, .47, "C"), (.53, .47, "D"))
panel_labels(root, labels)
normalize_axes(root)
image_name = "simulation." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def fig3(args):
"""
%prog fig3 chrA02,A02,C2,chrC02 chr.sizes all.bed data
Napus Figure 3 displays alignments between quartet chromosomes, inset
with read histograms.
"""
from jcvi.formats.bed import Bed
p = OptionParser(fig3.__doc__)
p.add_option("--gauge_step",
default=10000000,
type="int",
help="Step size for the base scale")
opts, args, iopts = p.set_image_options(args, figsize="12x9")
if len(args) != 4:
sys.exit(not p.print_help())
chrs, sizes, bedfile, datadir = args
gauge_step = opts.gauge_step
diverge = iopts.diverge
rr, gg = diverge
chrs = [[x] for x in chrs.split(",")]
sizes = Sizes(sizes).mapping
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
chr_sizes, chr_sum_sizes, ratio = calc_ratio(chrs, sizes)
# Synteny panel
seqidsfile = make_seqids(chrs)
klayout = make_layout(chrs, chr_sum_sizes, ratio, template_f3a, shift=.05)
height = .07
r = height / 4
K = Karyotype(fig,
root,
seqidsfile,
klayout,
gap=gap,
height=height,
lw=2,
generank=False,
sizes=sizes,
heightpad=r,
roundrect=True,
plot_label=False)
# Chromosome labels
for kl in K.layout:
if kl.empty:
continue
lx, ly = kl.xstart, kl.y
if lx < .11:
lx += .1
ly += .06
label = kl.label
root.text(lx - .015, ly, label, fontsize=15, ha="right", va="center")
# Inset with datafiles
datafiles = ("chrA02.bzh.forxmgr", "parent.A02.per10kb.forxmgr",
"parent.C2.per10kb.forxmgr", "chrC02.bzh.forxmgr")
datafiles = [op.join(datadir, x) for x in datafiles]
tracks = K.tracks
hlfile = op.join(datadir, "bzh.regions.forhaibao")
xy_axes = []
for t, datafile in zip(tracks, datafiles):
ax = make_affix_axis(fig, t, -r, height=2 * r)
xy_axes.append(ax)
chr = t.seqids[0]
xy = XYtrack(ax, datafile, color="lightslategray")
start, end = 0, t.total
xy.interpolate(end)
xy.cap(ymax=40)
xy.import_hlfile(hlfile, chr, diverge=diverge)
xy.draw()
ax.set_xlim(start, end)
gauge_ax = make_affix_axis(fig, t, -r)
adjust_spines(gauge_ax, ["bottom"])
setup_gauge_ax(gauge_ax, start, end, gauge_step)
# Converted gene tracks
ax_Ar = make_affix_axis(fig, tracks[1], r, height=r / 2)
ax_Co = make_affix_axis(fig, tracks[2], r, height=r / 2)
order = Bed(bedfile).order
for asterisk in (False, True):
conversion_track(order,
"data/Genes.Converted.seuil.0.6.AtoC.txt",
0,
"A02",
ax_Ar,
rr,
asterisk=asterisk)
conversion_track(order,
"data/Genes.Converted.seuil.0.6.AtoC.txt",
1,
"C2",
ax_Co,
gg,
asterisk=asterisk)
conversion_track(order,
"data/Genes.Converted.seuil.0.6.CtoA.txt",
0,
"A02",
ax_Ar,
gg,
ypos=1,
asterisk=asterisk)
conversion_track(order,
"data/Genes.Converted.seuil.0.6.CtoA.txt",
1,
"C2",
ax_Co,
rr,
ypos=1,
asterisk=asterisk)
Ar, Co = xy_axes[1:3]
annotations = ((Ar, "Bra028920 Bra028897", "center",
"1DAn2+"), (Ar, "Bra020081 Bra020171", "right", "2DAn2+"),
(Ar, "Bra020218 Bra020286", "left",
"3DAn2+"), (Ar, "Bra008143 Bra008167", "left", "4DAn2-"),
(Ar, "Bra029317 Bra029251", "right",
"5DAn2+ (GSL)"), (Co, "Bo2g001000 Bo2g001300", "left",
"1DCn2-"), (Co, "Bo2g018560 Bo2g023700",
"right", "2DCn2-"),
(Co, "Bo2g024450 Bo2g025390", "left",
"3DCn2-"), (Co, "Bo2g081060 Bo2g082340", "left", "4DCn2+"),
(Co, "Bo2g161510 Bo2g164260", "right", "5DCn2-"))
for ax, genes, ha, label in annotations:
g1, g2 = genes.split()
x1, x2 = order[g1][1].start, order[g2][1].start
if ha == "center":
x = (x1 + x2) / 2 * .8
elif ha == "left":
x = x2
else:
x = x1
label = r"\textit{{{0}}}".format(label)
color = rr if "+" in label else gg
ax.text(x, 30, label, color=color, fontsize=9, ha=ha, va="center")
ax_Ar.set_xlim(0, tracks[1].total)
ax_Ar.set_ylim(-1, 1)
ax_Co.set_xlim(0, tracks[2].total)
ax_Co.set_ylim(-1, 1)
# Plot coverage in resequencing lines
gstep = 5000000
order = "swede,kale,h165,yudal,aviso,abu,bristol".split(",")
labels_dict = {"h165": "Resynthesized (H165)", "abu": "Aburamasari"}
hlsuffix = "regions.forhaibao"
chr1, chr2 = "chrA02", "chrC02"
t1, t2 = tracks[0], tracks[-1]
s1, s2 = sizes[chr1], sizes[chr2]
canvas1 = (t1.xstart, .75, t1.xend - t1.xstart, .2)
c = Coverage(fig,
root,
canvas1,
chr1, (0, s1),
datadir,
order=order,
gauge=None,
plot_chr_label=False,
gauge_step=gstep,
palette="gray",
cap=40,
hlsuffix=hlsuffix,
labels_dict=labels_dict,
diverge=diverge)
yys = c.yys
x1, x2 = .37, .72
tip = .02
annotations = ((x1, yys[2] + .3 * tip, tip, tip / 2,
"FLC"), (x1, yys[3] + .6 * tip, tip, tip / 2, "FLC"),
(x1, yys[5] + .6 * tip, tip, tip / 2,
"FLC"), (x2, yys[0] + .9 * tip, -1.2 * tip, 0, "GSL"),
(x2, yys[4] + .9 * tip, -1.2 * tip, 0,
"GSL"), (x2, yys[6] + .9 * tip, -1.2 * tip, 0, "GSL"))
arrowprops = dict(facecolor='black',
shrink=.05,
frac=.5,
width=1,
headwidth=4)
for x, y, dx, dy, label in annotations:
label = r"\textit{{{0}}}".format(label)
root.annotate(label,
xy=(x, y),
xytext=(x + dx, y + dy),
arrowprops=arrowprops,
color=rr,
fontsize=9,
ha="center",
va="center")
canvas2 = (t2.xstart, .05, t2.xend - t2.xstart, .2)
Coverage(fig,
root,
canvas2,
chr2, (0, s2),
datadir,
order=order,
gauge=None,
plot_chr_label=False,
gauge_step=gstep,
palette="gray",
cap=40,
hlsuffix=hlsuffix,
labels_dict=labels_dict,
diverge=diverge)
pad = .03
labels = ((.1, .67, "A"), (t1.xstart - 3 * pad, .95 + pad, "B"),
(t2.xstart - 3 * pad, .25 + pad, "C"))
panel_labels(root, labels)
normalize_axes(root)
image_name = "napus-fig3." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def fig3(args):
"""
%prog fig3 chrA02,A02,C2,chrC02 chr.sizes all.bed data
Napus Figure 3 displays alignments between quartet chromosomes, inset
with read histograms.
"""
from jcvi.formats.bed import Bed
p = OptionParser(fig3.__doc__)
p.add_option("--gauge_step", default=10000000, type="int",
help="Step size for the base scale")
opts, args, iopts = p.set_image_options(args, figsize="12x9")
if len(args) != 4:
sys.exit(not p.print_help())
chrs, sizes, bedfile, datadir = args
gauge_step = opts.gauge_step
diverge = iopts.diverge
rr, gg = diverge
chrs = [[x] for x in chrs.split(",")]
sizes = Sizes(sizes).mapping
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
chr_sizes, chr_sum_sizes, ratio = calc_ratio(chrs, sizes)
# Synteny panel
seqidsfile = make_seqids(chrs)
klayout = make_layout(chrs, chr_sum_sizes, ratio, template_f3a, shift=.05)
height = .07
r = height / 4
K = Karyotype(fig, root, seqidsfile, klayout, gap=gap,
height=height, lw=2, generank=False, sizes=sizes,
heightpad=r, roundrect=True, plot_label=False)
# Chromosome labels
for kl in K.layout:
if kl.empty:
continue
lx, ly = kl.xstart, kl.y
if lx < .11:
lx += .1
ly += .06
label = kl.label
root.text(lx - .015, ly, label, fontsize=15,
ha="right", va="center")
# Inset with datafiles
datafiles = ("chrA02.bzh.forxmgr", "parent.A02.per10kb.forxmgr",
"parent.C2.per10kb.forxmgr", "chrC02.bzh.forxmgr")
datafiles = [op.join(datadir, x) for x in datafiles]
tracks = K.tracks
hlfile = op.join(datadir, "bzh.regions.forhaibao")
xy_axes = []
for t, datafile in zip(tracks, datafiles):
ax = make_affix_axis(fig, t, -r, height=2 * r)
xy_axes.append(ax)
chr = t.seqids[0]
xy = XYtrack(ax, datafile, color="lightslategray")
start, end = 0, t.total
xy.interpolate(end)
xy.cap(ymax=40)
xy.import_hlfile(hlfile, chr, diverge=diverge)
xy.draw()
ax.set_xlim(start, end)
gauge_ax = make_affix_axis(fig, t, -r)
adjust_spines(gauge_ax, ["bottom"])
setup_gauge_ax(gauge_ax, start, end, gauge_step)
# Converted gene tracks
ax_Ar = make_affix_axis(fig, tracks[1], r, height=r/2)
ax_Co = make_affix_axis(fig, tracks[2], r, height=r/2)
order = Bed(bedfile).order
for asterisk in (False, True):
conversion_track(order, "data/Genes.Converted.seuil.0.6.AtoC.txt",
0, "A02", ax_Ar, rr, asterisk=asterisk)
conversion_track(order, "data/Genes.Converted.seuil.0.6.AtoC.txt",
1, "C2", ax_Co, gg, asterisk=asterisk)
conversion_track(order, "data/Genes.Converted.seuil.0.6.CtoA.txt",
0, "A02", ax_Ar, gg, ypos=1, asterisk=asterisk)
conversion_track(order, "data/Genes.Converted.seuil.0.6.CtoA.txt",
1, "C2", ax_Co, rr, ypos=1, asterisk=asterisk)
Ar, Co = xy_axes[1:3]
annotations = ((Ar, "Bra028920 Bra028897", "center", "1DAn2+"),
(Ar, "Bra020081 Bra020171", "right", "2DAn2+"),
(Ar, "Bra020218 Bra020286", "left", "3DAn2+"),
(Ar, "Bra008143 Bra008167", "left", "4DAn2-"),
(Ar, "Bra029317 Bra029251", "right", "5DAn2+ (GSL)"),
(Co, "Bo2g001000 Bo2g001300", "left", "1DCn2-"),
(Co, "Bo2g018560 Bo2g023700", "right", "2DCn2-"),
(Co, "Bo2g024450 Bo2g025390", "left", "3DCn2-"),
(Co, "Bo2g081060 Bo2g082340", "left", "4DCn2+"),
(Co, "Bo2g161510 Bo2g164260", "right", "5DCn2-"))
for ax, genes, ha, label in annotations:
g1, g2 = genes.split()
x1, x2 = order[g1][1].start, order[g2][1].start
if ha == "center":
x = (x1 + x2) / 2 * .8
elif ha == "left":
x = x2
else:
x = x1
label = r"\textit{{{0}}}".format(label)
color = rr if "+" in label else gg
ax.text(x, 30, label, color=color, fontsize=9, ha=ha, va="center")
ax_Ar.set_xlim(0, tracks[1].total)
ax_Ar.set_ylim(-1, 1)
ax_Co.set_xlim(0, tracks[2].total)
ax_Co.set_ylim(-1, 1)
# Plot coverage in resequencing lines
gstep = 5000000
order = "swede,kale,h165,yudal,aviso,abu,bristol".split(",")
labels_dict = {"h165": "Resynthesized (H165)", "abu": "Aburamasari"}
hlsuffix = "regions.forhaibao"
chr1, chr2 = "chrA02", "chrC02"
t1, t2 = tracks[0], tracks[-1]
s1, s2 = sizes[chr1], sizes[chr2]
canvas1 = (t1.xstart, .75, t1.xend - t1.xstart, .2)
c = Coverage(fig, root, canvas1, chr1, (0, s1), datadir,
order=order, gauge=None, plot_chr_label=False,
gauge_step=gstep, palette="gray",
cap=40, hlsuffix=hlsuffix, labels_dict=labels_dict,
diverge=diverge)
yys = c.yys
x1, x2 = .37, .72
tip = .02
annotations = ((x1, yys[2] + .3 * tip, tip, tip / 2, "FLC"),
(x1, yys[3] + .6 * tip, tip, tip / 2, "FLC"),
(x1, yys[5] + .6 * tip, tip, tip / 2, "FLC"),
(x2, yys[0] + .9 * tip, -1.2 * tip, 0, "GSL"),
(x2, yys[4] + .9 * tip, -1.2 * tip, 0, "GSL"),
(x2, yys[6] + .9 * tip, -1.2 * tip, 0, "GSL"))
arrowprops=dict(facecolor='black', shrink=.05, frac=.5,
width=1, headwidth=4)
for x, y, dx, dy, label in annotations:
label = r"\textit{{{0}}}".format(label)
root.annotate(label, xy=(x, y), xytext=(x + dx, y + dy),
arrowprops=arrowprops, color=rr, fontsize=9,
ha="center", va="center")
canvas2 = (t2.xstart, .05, t2.xend - t2.xstart, .2)
Coverage(fig, root, canvas2, chr2, (0, s2), datadir,
order=order, gauge=None, plot_chr_label=False,
gauge_step=gstep, palette="gray",
cap=40, hlsuffix=hlsuffix, labels_dict=labels_dict,
diverge=diverge)
pad = .03
labels = ((.1, .67, "A"), (t1.xstart - 3 * pad, .95 + pad, "B"),
(t2.xstart - 3 * pad, .25 + pad, "C"))
panel_labels(root, labels)
normalize_axes(root)
image_name = "napus-fig3." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def compare(args):
"""
%prog compare Evaluation.csv
Compare performances of various variant callers on simulated STR datasets.
"""
p = OptionParser(__doc__)
opts, args, iopts = p.set_image_options(args, figsize="15x5")
if len(args) != 1:
sys.exit(not p.print_help())
datafile, = args
pf = datafile.rsplit(".", 1)[0]
fig, (ax1, ax2, ax3) = plt.subplots(ncols=3,
nrows=1,
figsize=(iopts.w, iopts.h))
plt.tight_layout(pad=2)
# Huntington risk allele
infected_thr = 40
ref_thr = 19
# ax1: Multiple callers at lower range
df = pd.read_csv("Evaluation.csv")
truth = df["Truth"]
ax1.plot(truth, df["Manta"], 'bx-')
ax1.plot(truth, df["Isaac"], 'yo-')
ax1.plot(truth, df["GATK"], 'md-')
ax1.plot(truth, df["lobSTR"], 'c+-')
ax1.plot(truth, truth, 'k--') # to show diagonal
bbox = {'facecolor': 'tomato', 'alpha': .2, 'ec': 'w'}
pad = 2
ax1.axhline(infected_thr, color='tomato')
ax1.text(max(truth) - pad,
infected_thr + pad,
'Risk threshold',
bbox=bbox,
ha="right")
ax1.axhline(ref_thr, color='tomato')
ax1.text(max(truth) - pad,
ref_thr - pad,
'Reference repeat count',
bbox=bbox,
ha="right",
va="top")
ax1.set_xlabel(r'Num of CAG repeats inserted ($\mathit{h}$)')
ax1.set_ylabel('Num of CAG repeats called')
ax1.set_title(r'Simulated haploid $\mathit{h}$')
ax1.legend(['Manta', 'Isaac', 'GATK', 'lobSTR', 'Truth'], loc='best')
max_insert = 120
# ax2: lobSTR vs TREDPARSE with haploid model
lobstr_results = parse_results("lobstr_results_homo.txt")
tredparse_results = parse_results("tredparse_results_homo.txt")
truth = range(10, max_insert + 1)
lx, ly = zip(*lobstr_results)
tx, ty = zip(*tredparse_results)
ax2.plot(lx, ly, 'c+-')
ax2.plot(tx, ty, 'gx-')
ax2.plot(truth, truth, 'k--')
ax2.set_xlabel(r'Num of CAG repeats inserted ($\mathit{h}$)')
ax2.set_ylabel('Num of CAG repeats called')
ax2.set_title(r'Simulated haploid $\mathit{h}$')
ax2.legend(['lobSTR', 'TREDPARSE', 'Truth'], loc='best')
pad *= 2
ax2.axhline(infected_thr, color='tomato')
ax2.text(max(truth) - pad,
infected_thr + pad,
'Risk threshold',
bbox=bbox,
ha="right")
ax2.set_xlim(10, max_insert)
# ax3: lobSTR vs TREDPARSE with haploid model
lobstr_results = parse_results("lobstr_results_het.txt", exclude=20)
tredparse_results = parse_results("tredparse_results_het.txt", exclude=20)
truth = range(10, max_insert + 1)
lx, ly = zip(*lobstr_results)
tx, ty = zip(*tredparse_results)
ax3.plot(lx, ly, 'c+-')
ax3.plot(tx, ty, 'gx-')
ax3.plot(truth, truth, 'k--')
ax3.set_xlabel(r'Num of CAG repeats inserted ($\mathit{h}$)')
ax3.set_ylabel('Num of CAG repeats called')
ax3.set_title(r'Simulated diploid $\mathit{20/h}$')
ax3.legend(['lobSTR', 'TREDPARSE', 'Truth'], loc='best')
ax3.axhline(infected_thr, color='tomato')
ax3.text(max(truth) - pad,
infected_thr + pad,
'Risk threshold',
bbox=bbox,
ha="right")
ax3.set_xlim(10, max_insert)
root = fig.add_axes([0, 0, 1, 1])
pad = .03
panel_labels(root, ((pad / 2, 1 - pad, "A"), (1 / 3., 1 - pad, "B"),
(2 / 3., 1 - pad, "C")))
normalize_axes(root)
image_name = pf + "." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def compare3(args):
"""
%prog compare3
Compare performances of various variant callers on simulated STR datasets.
This compares the power of various evidence types.
"""
p = OptionParser(compare3.__doc__)
p.add_option('--maxinsert',
default=300,
type="int",
help="Maximum number of repeats")
add_simulate_options(p)
opts, args, iopts = p.set_image_options(args, figsize="10x10")
if len(args) != 0:
sys.exit(not p.print_help())
max_insert = opts.maxinsert
fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(ncols=2,
nrows=2,
figsize=(iopts.w, iopts.h))
plt.tight_layout(pad=3)
color = "lightslategray"
# ax1: Spanning
tredparse_results = parse_results("tredparse_results_het-spanning.txt")
title = SIMULATED_DIPLOID + "( Sub-model 1: Spanning reads)"
plot_compare(ax1,
title,
tredparse_results,
None,
color=color,
max_insert=max_insert,
risk=False)
# ax2: Partial
tredparse_results = parse_results("tredparse_results_het-partial.txt",
exclude=20)
title = SIMULATED_DIPLOID + " (Sub-model 2: Partial reads)"
plot_compare(ax2,
title,
tredparse_results,
None,
color=color,
max_insert=max_insert,
risk=False)
# ax3: Repeat
tredparse_results = parse_results("tredparse_results_het-repeat.txt",
exclude=20)
# HACK (repeat reads won't work under 50)
tredparse_results = [x for x in tredparse_results if x[0] > 50]
title = SIMULATED_DIPLOID + " (Sub-model 3: Repeat-only reads)"
plot_compare(ax3,
title,
tredparse_results,
None,
color=color,
max_insert=max_insert,
risk=False)
# ax4: Pair
tredparse_results = parse_results("tredparse_results_het-pair.txt",
exclude=20)
title = SIMULATED_DIPLOID + " (Sub-model 4: Paired-end reads)"
plot_compare(ax4,
title,
tredparse_results,
None,
color=color,
max_insert=max_insert,
risk=False)
for ax in (ax1, ax2, ax3, ax4):
ax.set_xlim(0, max_insert)
ax.set_ylim(0, max_insert)
root = fig.add_axes([0, 0, 1, 1])
pad = .03
panel_labels(root, ((pad / 2, 1 - pad, "A"), (1 / 2., 1 - pad, "B"),
(pad / 2, 1 / 2., "C"), (1 / 2., 1 / 2., "D")))
normalize_axes(root)
image_name = "tredparse." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def compare4(args):
"""
%prog compare4
Compare performances of various variant callers on simulated STR datasets.
Adds coverage comparisons as panel C and D.
"""
p = OptionParser(compare4.__doc__)
p.add_option('--maxinsert',
default=300,
type="int",
help="Maximum number of repeats")
add_simulate_options(p)
opts, args, iopts = p.set_image_options(args, figsize="10x10")
if len(args) != 0:
sys.exit(not p.print_help())
depth = opts.depth
max_insert = opts.maxinsert
fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(ncols=2,
nrows=2,
figsize=(iopts.w, iopts.h))
plt.tight_layout(pad=3)
# ax1: lobSTR vs TREDPARSE with haploid model
lobstr_results = parse_results("lobstr_results_homo-20x-150bp-500bp.txt")
tredparse_results = parse_results(
"tredparse_results_homo-20x-150bp-500bp.txt")
title = SIMULATED_HAPLOID + r" ($Depth=%s\times)" % depth
plot_compare(ax1,
title,
tredparse_results,
lobstr_results,
max_insert=max_insert)
# ax2: lobSTR vs TREDPARSE with diploid model (depth=20x)
lobstr_results = parse_results("lobstr_results_het-20x-150bp-500bp.txt",
exclude=20)
tredparse_results = parse_results(
"tredparse_results_het-20x-150bp-500bp.txt", exclude=20)
title = SIMULATED_DIPLOID + r" ($Depth=%s\times$)" % depth
plot_compare(ax2,
title,
tredparse_results,
lobstr_results,
max_insert=max_insert)
# ax3: lobSTR vs TREDPARSE with diploid model (depth=5x)
lobstr_results = parse_results("lobstr_results_het-5x-150bp-500bp.txt",
exclude=20)
tredparse_results = parse_results(
"tredparse_results_het-5x-150bp-500bp.txt", exclude=20)
title = SIMULATED_DIPLOID + r" ($Depth=%s\times$)" % 5
plot_compare(ax3,
title,
tredparse_results,
lobstr_results,
max_insert=max_insert)
# ax4: lobSTR vs TREDPARSE with diploid model (depth=80x)
lobstr_results = parse_results("lobstr_results_het-80x-150bp-500bp.txt",
exclude=20)
tredparse_results = parse_results(
"tredparse_results_het-80x-150bp-500bp.txt", exclude=20)
title = SIMULATED_DIPLOID + r" ($Depth=%s\times$)" % 80
plot_compare(ax4,
title,
tredparse_results,
lobstr_results,
max_insert=max_insert)
for ax in (ax1, ax2, ax3, ax4):
ax.set_xlim(0, max_insert)
ax.set_ylim(0, max_insert)
root = fig.add_axes([0, 0, 1, 1])
pad = .03
panel_labels(root, ((pad / 2, 1 - pad, "A"), (1 / 2., 1 - pad, "B"),
(pad / 2, 1 / 2., "C"), (1 / 2., 1 / 2., "D")))
normalize_axes(root)
image_name = "tredparse." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def lms(args):
"""
%prog lms
ALLMAPS cartoon to illustrate LMS metric.
"""
from random import randint
from jcvi.graphics.chromosome import HorizontalChromosome
p = OptionParser(lms.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="6x6", dpi=300)
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
# Panel A
w, h = 0.7, 0.35
ax = fig.add_axes([0.15, 0.6, w, h])
xdata = [x + randint(-3, 3) for x in range(10, 110, 10)]
ydata = [x + randint(-3, 3) for x in range(10, 110, 10)]
ydata[3:7] = ydata[3:7][::-1]
xydata = zip(xdata, ydata)
lis = xydata[:3] + [xydata[4]] + xydata[7:]
lds = xydata[3:7]
xlis, ylis = zip(*lis)
xlds, ylds = zip(*lds)
ax.plot(
xlis,
ylis,
"r-",
lw=12,
alpha=0.3,
solid_capstyle="round",
solid_joinstyle="round",
)
ax.plot(
xlds,
ylds,
"g-",
lw=12,
alpha=0.3,
solid_capstyle="round",
solid_joinstyle="round",
)
ax.plot(xdata, ydata, "k.", mec="k", mfc="w", mew=3, ms=12)
HorizontalChromosome(root, 0.15, 0.15 + w, 0.57, height=0.02, lw=2)
root.text(0.15 + w / 2, 0.55, "Chromosome location (bp)", ha="center", va="top")
ax.text(80, 30, "LIS = 7", color="r", ha="center", va="center")
ax.text(80, 20, "LDS = 4", color="g", ha="center", va="center")
ax.text(80, 10, "LMS = $max$(LIS, LDS) = 7", ha="center", va="center")
normalize_lms_axis(ax, xlim=110, ylim=110)
# Panel B
w = 0.37
p = (0, 45, 75, 110)
ax = fig.add_axes([0.1, 0.12, w, h])
xdata = [x for x in range(10, 110, 10)]
ydata = ydata_orig = [x for x in range(10, 110, 10)]
ydata = ydata[:4] + ydata[7:] + ydata[4:7][::-1]
xydata = zip(xdata, ydata)
lis = xydata[:7]
xlis, ylis = zip(*lis)
ax.plot(
xlis,
ylis,
"r-",
lw=12,
alpha=0.3,
solid_capstyle="round",
solid_joinstyle="round",
)
ax.plot(xdata, ydata, "k.", mec="k", mfc="w", mew=3, ms=12)
ax.vlines(p, 0, 110, colors="beige", lw=3)
normalize_lms_axis(ax, xlim=110, ylim=110)
patch = [0.1 + w * x / 110.0 for x in p]
HorizontalChromosome(root, 0.1, 0.1 + w, 0.09, patch=patch, height=0.02, lw=2)
scaffolds = ("a", "b", "c")
for i, s in enumerate(scaffolds):
xx = (patch[i] + patch[i + 1]) / 2
root.text(xx, 0.09, s, va="center", ha="center")
root.text(0.1 + w / 2, 0.04, "LMS($a||b||c$) = 7", ha="center")
# Panel C
ax = fig.add_axes([0.6, 0.12, w, h])
patch = [0.6 + w * x / 110.0 for x in p]
ydata = ydata_orig
ax.plot(
xdata,
ydata,
"r-",
lw=12,
alpha=0.3,
solid_capstyle="round",
solid_joinstyle="round",
)
ax.plot(xdata, ydata, "k.", mec="k", mfc="w", mew=3, ms=12)
ax.vlines(p, [0], [110], colors="beige", lw=3)
normalize_lms_axis(ax, xlim=110, ylim=110)
HorizontalChromosome(root, 0.6, 0.6 + w, 0.09, patch=patch, height=0.02, lw=2)
scaffolds = ("a", "-c", "b")
for i, s in enumerate(scaffolds):
xx = (patch[i] + patch[i + 1]) / 2
root.text(xx, 0.09, s, va="center", ha="center")
root.text(0.6 + w / 2, 0.04, "LMS($a||-c||b$) = 10", ha="center")
labels = ((0.05, 0.95, "A"), (0.05, 0.48, "B"), (0.55, 0.48, "C"))
panel_labels(root, labels)
normalize_axes(root)
pf = "lms"
image_name = pf + "." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def estimategaps(args):
"""
%prog estimategaps JM-4 chr1 JMMale-1
Illustrate ALLMAPS gap estimation algorithm.
"""
p = OptionParser(estimategaps.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="6x6", dpi=300)
if len(args) != 3:
sys.exit(not p.print_help())
pf, seqid, mlg = args
bedfile = pf + ".lifted.bed"
agpfile = pf + ".agp"
function = lambda x: x.cm
cc = Map(bedfile, scaffold_info=True, function=function)
agp = AGP(agpfile)
g = GapEstimator(cc, agp, seqid, mlg, function=function)
pp, chrsize, mlgsize = g.pp, g.chrsize, g.mlgsize
spl, spld = g.spl, g.spld
g.compute_all_gaps(verbose=False)
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
# Panel A
xstart, ystart = .15, .65
w, h = .7, .3
t = np.linspace(0, chrsize, 1000)
ax = fig.add_axes([xstart, ystart, w, h])
mx, my = zip(*g.scatter_data)
rho = spearmanr(mx, my)
dsg = "g"
ax.vlines(pp, 0, mlgsize, colors="beige")
ax.plot(mx, my, ".", color=set2[3])
ax.plot(t, spl(t), "-", color=dsg)
ax.text(.05, .95, mlg, va="top", transform=ax.transAxes)
normalize_lms_axis(ax, xlim=chrsize, ylim=mlgsize,
ylabel="Genetic distance (cM)")
if rho < 0:
ax.invert_yaxis()
# Panel B
ystart -= .28
h = .25
ax = fig.add_axes([xstart, ystart, w, h])
ax.vlines(pp, 0, mlgsize, colors="beige")
ax.plot(t, spld(t), "-", lw=2, color=dsg)
ax.plot(pp, spld(pp), "o", mfc="w", mec=dsg, ms=5)
normalize_lms_axis(ax, xlim=chrsize, ylim=25 * 1e-6,
xfactor=1e-6, xlabel="Physical position (Mb)",
yfactor=1000000, ylabel="Recomb. rate\n(cM / Mb)")
# Panel C (specific to JMMale-1)
a, b = "scaffold_1076", "scaffold_861"
sizes = dict((x.component_id, (x.object_beg, x.object_end,
x.component_span, x.orientation)) \
for x in g.agp if not x.is_gap)
a_beg, a_end, asize, ao = sizes[a]
b_beg, b_end, bsize, bo = sizes[b]
gapsize = g.get_gapsize(a)
total_size = asize + gapsize + bsize
ratio = .6 / total_size
y = .16
pad = .03
pb_ratio = w / chrsize
# Zoom
lsg = "lightslategray"
root.plot((.15 + pb_ratio * a_beg, .2),
(ystart, ystart - .14), ":", color=lsg)
root.plot((.15 + pb_ratio * b_end, .3),
(ystart, ystart - .08), ":", color=lsg)
ends = []
for tag, size, marker, beg in zip((a, b), (asize, bsize), (49213, 81277),
(.2, .2 + (asize + gapsize) * ratio)):
end = beg + size * ratio
marker = beg + marker * ratio
ends.append((beg, end, marker))
root.plot((marker,), (y,), "o", color=lsg)
root.text((beg + end) / 2, y + pad, latex(tag),
ha="center", va="center")
HorizontalChromosome(root, beg, end, y, height=.025, fc='gainsboro')
begs, ends, markers = zip(*ends)
fontprop = dict(color=lsg, ha="center", va="center")
ypos = y + pad * 2
root.plot(markers, (ypos, ypos), "-", lw=2, color=lsg)
root.text(sum(markers) / 2, ypos + pad,
"Distance: 1.29cM $\Leftrightarrow$ 211,824bp (6.1 cM/Mb)", **fontprop)
ypos = y - pad
xx = markers[0], ends[0]
root.plot(xx, (ypos, ypos), "-", lw=2, color=lsg)
root.text(sum(xx) / 2, ypos - pad, "34,115bp", **fontprop)
xx = markers[1], begs[1]
root.plot(xx, (ypos, ypos), "-", lw=2, color=lsg)
root.text(sum(xx) / 2, ypos - pad, "81,276bp", **fontprop)
root.plot((ends[0], begs[1]), (y, y), ":", lw=2, color=lsg)
root.text(sum(markers) / 2, ypos - 3 * pad, r"$\textit{Estimated gap size: 96,433bp}$",
color="r", ha="center", va="center")
labels = ((.05, .95, 'A'), (.05, .6, 'B'), (.05, .27, 'C'))
panel_labels(root, labels)
normalize_axes(root)
pf = "estimategaps"
image_name = pf + "." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def composite_ccn(df, size=(12, 8)):
"""Plot composite ccn figure"""
fig = plt.figure(1, size)
ax1 = plt.subplot2grid((2, 2), (0, 0))
ax2 = plt.subplot2grid((2, 2), (0, 1))
ax3 = plt.subplot2grid((2, 2), (1, 0))
ax4 = plt.subplot2grid((2, 2), (1, 1))
mf = df[df["hli_calc_gender"] == "Male"]
age_label = "Chronological age (yr)"
ax1.scatter(
mf["hli_calc_age_sample_taken"],
mf["ccn.chrX"],
s=10,
marker=".",
color="lightslategray",
)
ax1.set_ylim(0.8, 1.1)
plot_fit_line(ax1, mf["hli_calc_age_sample_taken"], mf["ccn.chrX"])
ax1.set_ylabel("ChrX copy number")
ax1.set_title("ChrX copy number in Male")
ax2.scatter(
mf["hli_calc_age_sample_taken"],
mf["ccn.chrY"],
s=10,
marker=".",
color="lightslategray",
)
plot_fit_line(ax2, mf["hli_calc_age_sample_taken"], mf["ccn.chrY"])
ax2.set_ylim(0.8, 1.1)
ax2.set_ylabel("ChrY copy number")
ax2.set_title("ChrY copy number in Male")
ax3.scatter(
df["hli_calc_age_sample_taken"],
df["ccn.chr1"],
s=10,
marker=".",
color="lightslategray",
)
plot_fit_line(ax3, df["hli_calc_age_sample_taken"], df["ccn.chr1"])
ax3.set_ylim(1.8, 2.1)
ax3.set_ylabel("Chr1 copy number")
ax3.set_title("Chr1 copy number")
ax4.scatter(
df["hli_calc_age_sample_taken"],
df["ccn.chrM"],
s=10,
marker=".",
color="lightslategray",
)
plot_fit_line(ax4, df["hli_calc_age_sample_taken"], df["ccn.chrM"])
ax4.set_ylim(0, 400)
ax4.set_ylabel("Mitochondria copy number")
ax4.set_title("Mitochondria copy number")
from matplotlib.lines import Line2D
for ax in (ax1, ax2, ax3, ax4):
ax.set_xlabel(age_label)
plt.tight_layout()
root = fig.add_axes((0, 0, 1, 1))
labels = ((0.02, 0.98, "A"), (0.52, 0.98, "B"), (0.02, 0.5, "C"),
(0.52, 0.5, "D"))
panel_labels(root, labels)
root.set_xlim(0, 1)
root.set_ylim(0, 1)
root.set_axis_off()
def composite(df, sameGenderMZ, sameGenderDZ, size=(16, 24)):
"""Embed both absdiff figures and heritability figures."""
fig = plt.figure(1, size)
ax1a = plt.subplot2grid((6, 4), (0, 0), rowspan=2, colspan=1)
ax2a = plt.subplot2grid((6, 4), (0, 1), rowspan=2, colspan=1)
ax3a = plt.subplot2grid((6, 4), (0, 2), rowspan=2, colspan=1)
ax4a = plt.subplot2grid((6, 4), (0, 3), rowspan=2, colspan=1)
ax1b = plt.subplot2grid((6, 4), (2, 0), rowspan=2, colspan=2)
ax2b = plt.subplot2grid((6, 4), (2, 2), rowspan=2, colspan=2)
ax3b = plt.subplot2grid((6, 4), (4, 0), rowspan=2, colspan=2)
ax4b = plt.subplot2grid((6, 4), (4, 2), rowspan=2, colspan=2)
# Telomeres
telomeres = extract_trait(df, "Sample name", "telomeres.Length")
mzTelomeres = extract_twin_values(sameGenderMZ, telomeres)
dzTelomeres = extract_twin_values(sameGenderDZ, telomeres)
plot_paired_values(ax1b, mzTelomeres, dzTelomeres, label="Telomere length")
plot_abs_diff(ax1a, mzTelomeres, dzTelomeres, label="Telomere length")
# CCNX
CCNX = extract_trait(df, "Sample name", "ccn.chrX")
mzCCNX = extract_twin_values(sameGenderMZ, CCNX, gender="Female")
dzCCNX = extract_twin_values(sameGenderDZ, CCNX, gender="Female")
dzCCNX = filter_low_values(dzCCNX, 1.75)
plot_paired_values(ax2b,
mzCCNX,
dzCCNX,
gender="Female only",
label="ChrX copy number")
plot_abs_diff(ax2a, mzCCNX, dzCCNX, label="ChrX copy number")
# CCNY
CCNY = extract_trait(df, "Sample name", "ccn.chrY")
mzCCNY = extract_twin_values(sameGenderMZ, CCNY, gender="Male")
dzCCNY = extract_twin_values(sameGenderDZ, CCNY, gender="Male")
dzCCNY = filter_low_values(dzCCNY, 0.75)
plot_paired_values(ax3b,
mzCCNY,
dzCCNY,
gender="Male only",
label="ChrY copy number")
plot_abs_diff(ax3a, mzCCNY, dzCCNY, label="ChrY copy number")
# CCNY
TRA = extract_trait(df, "Sample name", "TRA.PPM")
mzTRA = extract_twin_values(sameGenderMZ, TRA)
dzTRA = extract_twin_values(sameGenderDZ, TRA)
plot_paired_values(ax4b, mzTRA, dzTRA, label="TCR-$\\alpha$ deletions")
plot_abs_diff(ax4a, mzTRA, dzTRA, label="TCR-$\\alpha$ deletions")
plt.tight_layout()
root = fig.add_axes((0, 0, 1, 1))
# ABCD absdiff, EFGH heritability
labels = (
(0.03, 0.99, "A"),
(0.27, 0.99, "B"),
(0.53, 0.99, "C"),
(0.77, 0.99, "D"),
(0.03, 0.67, "E"),
(0.53, 0.67, "F"),
(0.03, 0.34, "G"),
(0.53, 0.34, "H"),
)
panel_labels(root, labels)
root.set_xlim(0, 1)
root.set_ylim(0, 1)
root.set_axis_off()
def cartoon(args):
"""
%prog synteny.py
Generate cartoon illustration of SynFind.
"""
p = OptionParser(cartoon.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="10x7")
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
# Panel A
A = CartoonRegion(41)
A.draw(root, .35, .85, strip=False, color=False)
x1, x2 = A.x1, A.x2
lsg = "lightslategray"
pad = .01
xc, yc = .35, .88
arrowlen = x2 - xc - pad
arrowprops = dict(length_includes_head=True, width=.01, fc=lsg, lw=0,
head_length=arrowlen * .15, head_width=.03)
p = FancyArrow(xc - pad, yc, -arrowlen, 0, shape="left", **arrowprops)
root.add_patch(p)
p = FancyArrow(xc + pad, yc, arrowlen, 0, shape="right", **arrowprops)
root.add_patch(p)
yt = yc + 4 * pad
root.text((x1 + xc) / 2, yt, "20 genes upstream", ha="center")
root.text((x2 + xc) / 2, yt, "20 genes downstream", ha="center")
root.plot((xc,), (yc,), "o", mfc='w', mec=lsg, mew=2, lw=2, color=lsg)
root.text(xc, yt, "Query gene", ha="center")
# Panel B
A.draw(root, .35, .7, strip=False)
RoundRect(root, (.07, .49), .56, .14, fc='y', alpha=.2)
a = deepcopy(A)
a.evolve(mode='S', target=10)
a.draw(root, .35, .6)
b = deepcopy(A)
b.evolve(mode='F', target=8)
b.draw(root, .35, .56)
c = deepcopy(A)
c.evolve(mode='G', target=6)
c.draw(root, .35, .52)
for x in (a, b, c):
root.text(.64, x.y, "Score={0}".format(x.nonwhites), va="center")
# Panel C
A.truncate_between_flankers()
a.truncate_between_flankers()
b.truncate_between_flankers()
c.truncate_between_flankers(target=6)
plot_diagram(root, .14, .2, A, a, "S", "syntenic")
plot_diagram(root, .37, .2, A, b, "F", "missing, with both flankers")
plot_diagram(root, .6, .2, A, c, "G", "missing, with one flanker")
labels = ((.04, .95, 'A'), (.04, .75, 'B'), (.04, .4, 'C'))
panel_labels(root, labels)
# Descriptions
xt = .85
desc = ("Extract neighborhood",
"of *window* size",
"Count gene pairs within *window*",
"Find regions above *score* cutoff",
"Identify flankers",
"Annotate syntelog class"
)
for yt, t in zip((.88, .84, .64, .6, .3, .26), desc):
root.text(xt, yt, markup(t), ha="center", va="center")
root.set_xlim(0, 1)
root.set_ylim(0, 1)
root.set_axis_off()
pf = "cartoon"
image_name = pf + "." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def plot(args):
"""
%prog plot input.bed seqid
Plot the matchings between the reconstructed pseudomolecules and the maps.
Two types of visualizations are available in one canvas:
1. Parallel axes, and matching markers are shown in connecting lines;
2. Scatter plot.
"""
from jcvi.graphics.base import plt, savefig, normalize_axes, \
set2, panel_labels
from jcvi.graphics.chromosome import Chromosome, GeneticMap, \
HorizontalChromosome
p = OptionParser(plot.__doc__)
add_allmaps_plot_options(p)
opts, args, iopts = p.set_image_options(args, figsize="10x6")
if len(args) != 2:
sys.exit(not p.print_help())
inputbed, seqid = args
pf = inputbed.rsplit(".", 1)[0]
bedfile = pf + ".lifted.bed"
agpfile = pf + ".agp"
weightsfile = opts.weightsfile
links = opts.links
function = get_function(opts.distance)
cc = Map(bedfile, function)
allseqids = cc.seqids
mapnames = cc.mapnames
weights = Weights(weightsfile, mapnames)
assert seqid in allseqids, "{0} not in {1}".format(seqid, allseqids)
s = Scaffold(seqid, cc)
mlgs = [k for k, v in s.mlg_counts.items() if v >= links]
mlgsizes = {}
for mlg in mlgs:
mm = cc.extract_mlg(mlg)
mlgsize = max(function(x) for x in mm)
mlgsizes[mlg] = mlgsize
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
ax1 = fig.add_axes([0, 0, .5, 1])
ax2 = fig.add_axes([.5, 0, .5, 1])
# Find the layout first
ystart, ystop = .9, .1
L = Layout(mlgsizes)
coords = L.coords
tip = .02
marker_pos = {}
# Palette
colors = dict((mapname, set2[i]) for i, mapname in enumerate(mapnames))
colors = dict((mlg, colors[mlg.split("-")[0]]) for mlg in mlgs)
rhos = {}
# Parallel coordinates
for mlg, (x, y1, y2) in coords.items():
mm = cc.extract_mlg(mlg)
markers = [(m.accn, function(m)) for m in mm] # exhaustive marker list
xy = [(m.pos, function(m)) for m in mm if m.seqid == seqid]
mx, my = zip(*xy)
rho = spearmanr(mx, my)
rhos[mlg] = rho
flip = rho < 0
g = GeneticMap(ax1, x, y1, y2, markers, tip=tip, flip=flip)
extra = -3 * tip if x < .5 else 3 * tip
ha = "right" if x < .5 else "left"
mapname = mlg.split("-")[0]
tlg = mlg.replace("_", ".") # Latex does not like underscore char
label = "{0} (w={1})".format(tlg, weights[mapname])
ax1.text(x + extra, (y1 + y2) / 2,
label,
color=colors[mlg],
ha=ha,
va="center",
rotation=90)
marker_pos.update(g.marker_pos)
agp = AGP(agpfile)
agp = [x for x in agp if x.object == seqid]
chrsize = max(x.object_end for x in agp)
# Pseudomolecules in the center
r = ystart - ystop
ratio = r / chrsize
f = lambda x: (ystart - ratio * x)
patchstart = [f(x.object_beg) for x in agp if not x.is_gap]
Chromosome(ax1, .5, ystart, ystop, width=2 * tip, patch=patchstart, lw=2)
label = "{0} ({1})".format(seqid, human_size(chrsize, precision=0))
ax1.text(.5, ystart + tip, label, ha="center")
scatter_data = defaultdict(list)
# Connecting lines
for b in s.markers:
marker_name = b.accn
if marker_name not in marker_pos:
continue
cx = .5
cy = f(b.pos)
mx = coords[b.mlg][0]
my = marker_pos[marker_name]
extra = -tip if mx < cx else tip
extra *= 1.25 # leave boundaries for aesthetic reasons
cx += extra
mx -= extra
ax1.plot((cx, mx), (cy, my), "-", color=colors[b.mlg])
scatter_data[b.mlg].append((b.pos, function(b)))
# Scatter plot, same data as parallel coordinates
xstart, xstop = sorted((ystart, ystop))
f = lambda x: (xstart + ratio * x)
pp = [x.object_beg for x in agp if not x.is_gap]
patchstart = [f(x) for x in pp]
HorizontalChromosome(ax2,
xstart,
xstop,
ystop,
height=2 * tip,
patch=patchstart,
lw=2)
gap = .03
ratio = (r - gap * len(mlgs) - tip) / sum(mlgsizes.values())
tlgs = []
for mlg, mlgsize in sorted(mlgsizes.items()):
height = ratio * mlgsize
ystart -= height
xx = .5 + xstart / 2
width = r / 2
color = colors[mlg]
ax = fig.add_axes([xx, ystart, width, height])
ypos = ystart + height / 2
ystart -= gap
sd = scatter_data[mlg]
xx, yy = zip(*sd)
ax.vlines(pp, 0, mlgsize, colors="beige")
ax.plot(xx, yy, ".", color=color)
rho = rhos[mlg]
ax.text(.5,
1 - .4 * gap / height,
r"$\rho$={0:.3f}".format(rho),
ha="center",
va="top",
transform=ax.transAxes,
color="gray")
tlg = mlg.replace("_", ".")
tlgs.append((tlg, ypos, color))
ax.set_xlim(0, chrsize)
ax.set_ylim(0, mlgsize)
ax.set_xticks([])
while height / len(ax.get_yticks()) < .03 and len(
ax.get_yticks()) >= 2:
ax.set_yticks(ax.get_yticks()[::2]) # Sparsify the ticks
yticklabels = [int(x) for x in ax.get_yticks()]
ax.set_yticklabels(yticklabels, family='Helvetica')
if rho < 0:
ax.invert_yaxis()
for i, (tlg, ypos, color) in enumerate(tlgs):
ha = "center"
if len(tlgs) > 4:
ha = "right" if i % 2 else "left"
root.text(.5, ypos, tlg, color=color, rotation=90, ha=ha, va="center")
if opts.panels:
labels = ((.04, .96, 'A'), (.48, .96, 'B'))
panel_labels(root, labels)
normalize_axes((ax1, ax2, root))
image_name = seqid + "." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
plt.close(fig)
def likelihood(args):
"""
%prog likelihood
Plot likelihood surface. Look for two files in the current folder:
- 100_100.log, haploid model
- 100_20.log, diploid model
"""
p = OptionParser(likelihood.__doc__)
opts, args, iopts = p.set_image_options(args,
figsize="10x5",
style="white",
cmap="coolwarm")
if len(args) != 0:
sys.exit(not p.print_help())
fig, (ax1, ax2) = plt.subplots(ncols=2,
nrows=1,
figsize=(iopts.w, iopts.h))
plt.tight_layout(pad=4)
# Haploid model
LL, CI_h1, CI_h2, MLE = parse_log("100_100.log")
data = []
for k, v in LL.items():
data.append((k[0], v))
data.sort()
x, y = zip(*data)
x = np.array(x)
curve, = ax1.plot(x, y, "-", color=lsg, lw=2)
ax1.set_title("Simulated haploid ($h^{truth}=100$)")
h_hat, max_LL = max(data, key=lambda x: x[-1])
_, min_LL = min(data, key=lambda x: x[-1])
ymin, ymax = ax1.get_ylim()
ax1.set_ylim([ymin, ymax + 30])
LL_label = "log(Likelihood)"
ax1.plot([h_hat, h_hat], [ymin, max_LL], ":", color=lsg, lw=2)
ax1.text(h_hat, max_LL + 10, r"$\hat{h}=93$", color=lsg)
ax1.set_xlabel(r"$h$")
ax1.set_ylabel(LL_label)
a, b = CI_h1
ci = ax1.fill_between(x, [ymin] * len(x),
y,
where=(x >= a) & (x <= b),
color=lsg,
alpha=.5)
ax1.legend([curve, ci], ["Likelihood curve", r'95$\%$ CI'], loc='best')
# Diploid model
LL, CI_h1, CI_h2, MLE = parse_log("100_20.log")
h_hat, max_LL = max(data, key=lambda x: x[-1])
_, min_LL = min(data, key=lambda x: x[-1])
data = np.ones((301, 301)) * min_LL
for k, v in LL.items():
a, b = k
data[a, b] = v
data[b, a] = v
data = mask_upper_triangle(data)
ax_imshow(ax2, data, opts.cmap, LL_label, 20, 104)
root = fig.add_axes([0, 0, 1, 1])
pad = .04
panel_labels(root, ((pad / 2, 1 - pad, "A"), (1 / 2., 1 - pad, "B")))
normalize_axes(root)
image_name = "likelihood." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def phylogeny(args):
"""
%prog phylogeny treefile ks.layout
Create a composite figure with (A) tree and (B) ks.
"""
from jcvi.graphics.tree import parse_tree, LeafInfoFile, WGDInfoFile, draw_tree
p = OptionParser(phylogeny.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="10x12")
(datafile, layoutfile) = args
logging.debug("Load tree file `{0}`".format(datafile))
t, hpd = parse_tree(datafile)
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
ax1 = fig.add_axes([0, 0.4, 1, 0.6])
ax2 = fig.add_axes([0.12, 0.065, 0.8, 0.3])
margin, rmargin = 0.1, 0.2 # Left and right margin
leafinfo = LeafInfoFile("leafinfo.csv").cache
wgdinfo = WGDInfoFile("wgdinfo.csv").cache
outgroup = "ginkgo"
# Panel A
draw_tree(
ax1,
t,
hpd=hpd,
margin=margin,
rmargin=rmargin,
supportcolor=None,
internal=False,
outgroup=outgroup,
reroot=False,
leafinfo=leafinfo,
wgdinfo=wgdinfo,
geoscale=True,
)
from jcvi.apps.ks import Layout, KsPlot, KsFile
# Panel B
ks_min = 0.0
ks_max = 3.0
bins = 60
fill = False
layout = Layout(layoutfile)
print(layout, file=sys.stderr)
kp = KsPlot(ax2, ks_max, bins, legendp="upper right")
for lo in layout:
data = KsFile(lo.ksfile)
data = [x.ng_ks for x in data]
data = [x for x in data if ks_min <= x <= ks_max]
kp.add_data(
data,
lo.components,
label=lo.label,
color=lo.color,
marker=lo.marker,
fill=fill,
fitted=False,
kde=True,
)
kp.draw(filename=None)
normalize_axes([root, ax1])
labels = ((0.05, 0.95, "A"), (0.05, 0.4, "B"))
panel_labels(root, labels)
image_name = "phylogeny.pdf"
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def lms(args):
"""
%prog lms
ALLMAPS cartoon to illustrate LMS metric.
"""
from random import randint
from jcvi.graphics.chromosome import HorizontalChromosome
p = OptionParser(lms.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="6x6", dpi=300)
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
# Panel A
w, h = .7, .35
ax = fig.add_axes([.15, .6, w, h])
xdata = [x + randint(-3, 3) for x in range(10, 110, 10)]
ydata = [x + randint(-3, 3) for x in range(10, 110, 10)]
ydata[3:7] = ydata[3:7][::-1]
xydata = zip(xdata, ydata)
lis = xydata[:3] + [xydata[4]] + xydata[7:]
lds = xydata[3:7]
xlis, ylis = zip(*lis)
xlds, ylds = zip(*lds)
ax.plot(xlis, ylis, "r-", lw=12, alpha=.3,
solid_capstyle="round", solid_joinstyle="round")
ax.plot(xlds, ylds, "g-", lw=12, alpha=.3,
solid_capstyle="round", solid_joinstyle="round")
ax.plot(xdata, ydata, "k.", mec="k", mfc="w", mew=3, ms=12)
HorizontalChromosome(root, .15, .15 + w, .57, height=.02, lw=2)
root.text(.15 + w / 2, .55, "Chromosome location (bp)", ha="center", va="top")
ax.text(80, 30, "LIS = 7", color="r", ha="center", va="center")
ax.text(80, 20, "LDS = 4", color="g", ha="center", va="center")
ax.text(80, 10, "LMS = $max$(LIS, LDS) = 7", ha="center", va="center")
normalize_lms_axis(ax)
# Panel B
w = .37
p = (0, 45, 75, 110)
ax = fig.add_axes([.1, .12, w, h])
xdata = [x for x in range(10, 110, 10)]
ydata = ydata_orig = [x for x in range(10, 110, 10)]
ydata = ydata[:4] + ydata[7:] + ydata[4:7][::-1]
xydata = zip(xdata, ydata)
lis = xydata[:7]
xlis, ylis = zip(*lis)
ax.plot(xlis, ylis, "r-", lw=12, alpha=.3,
solid_capstyle="round", solid_joinstyle="round")
ax.plot(xdata, ydata, "k.", mec="k", mfc="w", mew=3, ms=12)
ax.vlines(p, 0, 110, colors="beige", lw=3)
normalize_lms_axis(ax)
patch = [.1 + w * x / 110. for x in p]
HorizontalChromosome(root, .1, .1 + w, .09, patch=patch,
height=.02, lw=2)
scaffolds = ("a", "b", "c")
for i, s in enumerate(scaffolds):
xx = (patch[i] + patch[i + 1]) / 2
root.text(xx, .09, s, va="center", ha="center")
root.text(.1 + w / 2, .04, "LMS($a||b||c$) = 7", ha="center")
# Panel C
ax = fig.add_axes([.6, .12, w, h])
patch = [.6 + w * x / 110. for x in p]
ydata = ydata_orig
ax.plot(xdata, ydata, "r-", lw=12, alpha=.3,
solid_capstyle="round", solid_joinstyle="round")
ax.plot(xdata, ydata, "k.", mec="k", mfc="w", mew=3, ms=12)
ax.vlines(p, [0], [110], colors="beige", lw=3)
normalize_lms_axis(ax)
HorizontalChromosome(root, .6, .6 + w, .09, patch=patch,
height=.02, lw=2)
scaffolds = ("a", "-c", "b")
for i, s in enumerate(scaffolds):
xx = (patch[i] + patch[i + 1]) / 2
root.text(xx, .09, s, va="center", ha="center")
root.text(.6 + w / 2, .04, "LMS($a||-c||b$) = 10", ha="center")
labels = ((.05, .95, 'A'), (.05, .48, 'B'), (.55, .48, 'C'))
panel_labels(root, labels)
normalize_axes(root)
pf = "lms"
image_name = pf + "." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def composite_qc(df_orig, size=(16, 12)):
"""Plot composite QC figures"""
df = df_orig.rename(
columns={
"hli_calc_age_sample_taken": "Age",
"hli_calc_gender": "Gender",
"eth7_max": "Ethnicity",
"MeanCoverage": "Mean coverage",
"Chemistry": "Sequencing chemistry",
"Release Client": "Cohort",
})
fig = plt.figure(1, size)
ax1 = plt.subplot2grid((2, 7), (0, 0), rowspan=1, colspan=2)
ax2 = plt.subplot2grid((2, 7), (0, 2), rowspan=1, colspan=2)
ax3 = plt.subplot2grid((2, 7), (0, 4), rowspan=1, colspan=3)
ax4 = plt.subplot2grid((2, 7), (1, 0), rowspan=1, colspan=2)
ax5 = plt.subplot2grid((2, 7), (1, 2), rowspan=1, colspan=2)
ax6 = plt.subplot2grid((2, 7), (1, 4), rowspan=1, colspan=3)
sns.distplot(df["Age"].dropna(), kde=False, ax=ax1)
sns.countplot(x="Gender", data=df, ax=ax2)
sns.countplot(x="Ethnicity",
data=df,
ax=ax3,
order=df["Ethnicity"].value_counts().index)
sns.distplot(df["Mean coverage"].dropna(), kde=False, ax=ax4)
ax4.set_xlim(0, 100)
sns.countplot(x="Sequencing chemistry", data=df, ax=ax5)
sns.countplot(x="Cohort",
data=df,
ax=ax6,
order=df["Cohort"].value_counts().index)
# Anonymize the cohorts
cohorts = ax6.get_xticklabels()
newCohorts = []
for i, c in enumerate(cohorts):
if c.get_text() == "Spector":
c = "TwinsUK"
elif c.get_text() != "Health Nucleus":
c = "C{}".format(i + 1)
newCohorts.append(c)
ax6.set_xticklabels(newCohorts)
for ax in (ax6, ):
ax.set_xticklabels(ax.get_xticklabels(), ha="right", rotation=30)
for ax in (ax1, ax2, ax3, ax4, ax5, ax6):
ax.set_title(ax.get_xlabel())
ax.set_xlabel("")
plt.tight_layout()
root = fig.add_axes((0, 0, 1, 1))
labels = (
(0.02, 0.96, "A"),
(0.3, 0.96, "B"),
(0.6, 0.96, "C"),
(0.02, 0.52, "D"),
(0.3, 0.52, "E"),
(0.6, 0.52, "F"),
)
panel_labels(root, labels)
root.set_xlim(0, 1)
root.set_ylim(0, 1)
root.set_axis_off()
def composite_ccn(df, size=(12, 8)):
""" Plot composite ccn figure
"""
fig = plt.figure(1, size)
ax1 = plt.subplot2grid((2, 2), (0, 0))
ax2 = plt.subplot2grid((2, 2), (0, 1))
ax3 = plt.subplot2grid((2, 2), (1, 0))
ax4 = plt.subplot2grid((2, 2), (1, 1))
chemistry = ["V1", "V2", "V2.5", float("nan")]
colors = sns.color_palette("Set2", 8)
color_map = dict(zip(chemistry, colors))
mf = df[df["hli_calc_gender"] == "Male"]
age_label = "Chronological age (yr)"
ax1.scatter(mf["hli_calc_age_sample_taken"],
mf["ccn.chrX"],
s=10,
marker='.',
color='lightslategray')
ax1.set_ylim(0.8, 1.1)
plot_fit_line(ax1, mf["hli_calc_age_sample_taken"], mf["ccn.chrX"])
ax1.set_ylabel("ChrX copy number")
ax1.set_title("ChrX copy number in Male")
ax2.scatter(mf["hli_calc_age_sample_taken"],
mf["ccn.chrY"],
s=10,
marker='.',
color='lightslategray')
plot_fit_line(ax2, mf["hli_calc_age_sample_taken"], mf["ccn.chrY"])
ax2.set_ylim(0.8, 1.1)
ax2.set_ylabel("ChrY copy number")
ax2.set_title("ChrY copy number in Male")
ax3.scatter(df["hli_calc_age_sample_taken"],
df["ccn.chr1"],
s=10,
marker='.',
color='lightslategray')
plot_fit_line(ax3, df["hli_calc_age_sample_taken"], df["ccn.chr1"])
ax3.set_ylim(1.8, 2.1)
ax3.set_ylabel("Chr1 copy number")
ax3.set_title("Chr1 copy number")
ax4.scatter(df["hli_calc_age_sample_taken"],
df["ccn.chrM"],
s=10,
marker='.',
color='lightslategray')
plot_fit_line(ax4, df["hli_calc_age_sample_taken"], df["ccn.chrM"])
ax4.set_ylim(0, 400)
ax4.set_ylabel("Mitochondria copy number")
ax4.set_title("Mitochondria copy number")
from matplotlib.lines import Line2D
legend_elements = [Line2D([0], [0], marker='.', color='w', label=chem,
markerfacecolor=color) \
for (chem, color) in zip(chemistry, colors)[:3]]
for ax in (ax1, ax2, ax3, ax4):
ax.set_xlabel(age_label)
plt.tight_layout()
root = fig.add_axes((0, 0, 1, 1))
labels = ((.02, .98, "A"), (.52, .98, "B"), (.02, .5, "C"), (.52, .5, "D"))
panel_labels(root, labels)
root.set_xlim(0, 1)
root.set_ylim(0, 1)
root.set_axis_off()
def estimategaps(args):
"""
%prog estimategaps JM-4 chr1 JMMale-1
Illustrate ALLMAPS gap estimation algorithm.
"""
p = OptionParser(estimategaps.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="6x6", dpi=300)
if len(args) != 3:
sys.exit(not p.print_help())
pf, seqid, mlg = args
bedfile = pf + ".lifted.bed"
agpfile = pf + ".agp"
function = lambda x: x.cm
cc = Map(bedfile, scaffold_info=True, function=function)
agp = AGP(agpfile)
g = GapEstimator(cc, agp, seqid, mlg, function=function)
pp, chrsize, mlgsize = g.pp, g.chrsize, g.mlgsize
spl, spld = g.spl, g.spld
g.compute_all_gaps(verbose=False)
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
# Panel A
xstart, ystart = 0.15, 0.65
w, h = 0.7, 0.3
t = np.linspace(0, chrsize, 1000)
ax = fig.add_axes([xstart, ystart, w, h])
mx, my = zip(*g.scatter_data)
rho = spearmanr(mx, my)
dsg = "g"
ax.vlines(pp, 0, mlgsize, colors="beige")
ax.plot(mx, my, ".", color=set2[3])
ax.plot(t, spl(t), "-", color=dsg)
ax.text(0.05, 0.95, mlg, va="top", transform=ax.transAxes)
normalize_lms_axis(ax, xlim=chrsize, ylim=mlgsize, ylabel="Genetic distance (cM)")
if rho < 0:
ax.invert_yaxis()
# Panel B
ystart -= 0.28
h = 0.25
ax = fig.add_axes([xstart, ystart, w, h])
ax.vlines(pp, 0, mlgsize, colors="beige")
ax.plot(t, spld(t), "-", lw=2, color=dsg)
ax.plot(pp, spld(pp), "o", mfc="w", mec=dsg, ms=5)
normalize_lms_axis(
ax,
xlim=chrsize,
ylim=25 * 1e-6,
xfactor=1e-6,
xlabel="Physical position (Mb)",
yfactor=1000000,
ylabel="Recomb. rate\n(cM / Mb)",
)
ax.xaxis.grid(False)
# Panel C (specific to JMMale-1)
a, b = "scaffold_1076", "scaffold_861"
sizes = dict(
(x.component_id, (x.object_beg, x.object_end, x.component_span, x.orientation))
for x in g.agp
if not x.is_gap
)
a_beg, a_end, asize, ao = sizes[a]
b_beg, b_end, bsize, bo = sizes[b]
gapsize = g.get_gapsize(a)
total_size = asize + gapsize + bsize
ratio = 0.6 / total_size
y = 0.16
pad = 0.03
pb_ratio = w / chrsize
# Zoom
lsg = "lightslategray"
root.plot((0.15 + pb_ratio * a_beg, 0.2), (ystart, ystart - 0.14), ":", color=lsg)
root.plot((0.15 + pb_ratio * b_end, 0.3), (ystart, ystart - 0.08), ":", color=lsg)
ends = []
for tag, size, marker, beg in zip(
(a, b), (asize, bsize), (49213, 81277), (0.2, 0.2 + (asize + gapsize) * ratio)
):
end = beg + size * ratio
marker = beg + marker * ratio
ends.append((beg, end, marker))
root.plot((marker,), (y,), "o", color=lsg)
root.text((beg + end) / 2, y + pad, latex(tag), ha="center", va="center")
HorizontalChromosome(root, beg, end, y, height=0.025, fc="gainsboro")
begs, ends, markers = zip(*ends)
fontprop = dict(color=lsg, ha="center", va="center")
ypos = y + pad * 2
root.plot(markers, (ypos, ypos), "-", lw=2, color=lsg)
root.text(
sum(markers) / 2,
ypos + pad,
"Distance: 1.29cM $\Leftrightarrow$ 211,824bp (6.1 cM/Mb)",
**fontprop
)
ypos = y - pad
xx = markers[0], ends[0]
root.plot(xx, (ypos, ypos), "-", lw=2, color=lsg)
root.text(sum(xx) / 2, ypos - pad, "34,115bp", **fontprop)
xx = markers[1], begs[1]
root.plot(xx, (ypos, ypos), "-", lw=2, color=lsg)
root.text(sum(xx) / 2, ypos - pad, "81,276bp", **fontprop)
root.plot((ends[0], begs[1]), (y, y), ":", lw=2, color=lsg)
root.text(
sum(markers) / 2,
ypos - 3 * pad,
r"$\textit{Estimated gap size: 96,433bp}$",
color="r",
ha="center",
va="center",
)
labels = ((0.05, 0.95, "A"), (0.05, 0.6, "B"), (0.05, 0.27, "C"))
panel_labels(root, labels)
normalize_axes(root)
pf = "estimategaps"
image_name = pf + "." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def composite_ccn(df, size=(12, 8)):
""" Plot composite ccn figure
"""
fig = plt.figure(1, size)
ax1 = plt.subplot2grid((2, 2), (0, 0))
ax2 = plt.subplot2grid((2, 2), (0, 1))
ax3 = plt.subplot2grid((2, 2), (1, 0))
ax4 = plt.subplot2grid((2, 2), (1, 1))
chemistry = ["V1", "V2", "V2.5", float("nan")]
colors = sns.color_palette("Set2", 8)
color_map = dict(zip(chemistry, colors))
mf = df[df["hli_calc_gender"] == "Male"]
age_label = "Chronological age (yr)"
ax1.scatter(mf["hli_calc_age_sample_taken"], mf["ccn.chrX"],
s=10, marker='.',
color='lightslategray')
ax1.set_ylim(0.8, 1.1)
plot_fit_line(ax1, mf["hli_calc_age_sample_taken"], mf["ccn.chrX"])
ax1.set_ylabel("ChrX copy number")
ax1.set_title("ChrX copy number in Male")
ax2.scatter(mf["hli_calc_age_sample_taken"], mf["ccn.chrY"],
s=10, marker='.',
color='lightslategray')
plot_fit_line(ax2, mf["hli_calc_age_sample_taken"], mf["ccn.chrY"])
ax2.set_ylim(0.8, 1.1)
ax2.set_ylabel("ChrY copy number")
ax2.set_title("ChrY copy number in Male")
ax3.scatter(df["hli_calc_age_sample_taken"], df["ccn.chr1"],
s=10, marker='.',
color='lightslategray')
plot_fit_line(ax3, df["hli_calc_age_sample_taken"], df["ccn.chr1"])
ax3.set_ylim(1.8, 2.1)
ax3.set_ylabel("Chr1 copy number")
ax3.set_title("Chr1 copy number")
ax4.scatter(df["hli_calc_age_sample_taken"], df["ccn.chrM"],
s=10, marker='.',
color='lightslategray')
plot_fit_line(ax4, df["hli_calc_age_sample_taken"], df["ccn.chrM"])
ax4.set_ylim(0, 400)
ax4.set_ylabel("Mitochondria copy number")
ax4.set_title("Mitochondria copy number")
from matplotlib.lines import Line2D
legend_elements = [Line2D([0], [0], marker='.', color='w', label=chem,
markerfacecolor=color) \
for (chem, color) in zip(chemistry, colors)[:3]]
for ax in (ax1, ax2, ax3, ax4):
ax.set_xlabel(age_label)
plt.tight_layout()
root = fig.add_axes((0, 0, 1, 1))
labels = ((.02, .98, "A"),
(.52, .98, "B"),
(.02, .5, "C"),
(.52, .5, "D"))
panel_labels(root, labels)
root.set_xlim(0, 1)
root.set_ylim(0, 1)
root.set_axis_off()
def cartoon(args):
"""
%prog synteny.py
Generate cartoon illustration of SynFind.
"""
p = OptionParser(cartoon.__doc__)
opts, args, iopts = p.set_image_options(args, figsize="10x7")
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
# Panel A
A = CartoonRegion(41)
A.draw(root, .35, .85, strip=False, color=False)
x1, x2 = A.x1, A.x2
lsg = "lightslategray"
pad = .01
xc, yc = .35, .88
arrowlen = x2 - xc - pad
arrowprops = dict(length_includes_head=True, width=.01, fc=lsg, lw=0,
head_length=arrowlen * .15, head_width=.03)
p = FancyArrow(xc - pad, yc, -arrowlen, 0, shape="left", **arrowprops)
root.add_patch(p)
p = FancyArrow(xc + pad, yc, arrowlen, 0, shape="right", **arrowprops)
root.add_patch(p)
yt = yc + 4 * pad
root.text((x1 + xc) / 2, yt, "20 genes upstream", ha="center")
root.text((x2 + xc) / 2, yt, "20 genes downstream", ha="center")
root.plot((xc,), (yc,), "o", mfc='w', mec=lsg, mew=2, lw=2, color=lsg)
root.text(xc, yt, "Query gene", ha="center")
# Panel B
A.draw(root, .35, .7, strip=False)
RoundRect(root, (.07, .49), .56, .14, fc='y', alpha=.2)
a = deepcopy(A)
a.evolve(mode='S', target=10)
a.draw(root, .35, .6)
b = deepcopy(A)
b.evolve(mode='F', target=8)
b.draw(root, .35, .56)
c = deepcopy(A)
c.evolve(mode='G', target=6)
c.draw(root, .35, .52)
for x in (a, b, c):
root.text(.64, x.y, "Score={0}".format(x.nonwhites), va="center")
# Panel C
A.truncate_between_flankers()
a.truncate_between_flankers()
b.truncate_between_flankers()
c.truncate_between_flankers(target=6)
plot_diagram(root, .14, .2, A, a, "S", "syntenic")
plot_diagram(root, .37, .2, A, b, "F", "missing, with both flankers")
plot_diagram(root, .6, .2, A, c, "G", "missing, with one flanker")
labels = ((.04, .95, 'A'), (.04, .75, 'B'), (.04, .4, 'C'))
panel_labels(root, labels)
# Descriptions
xt = .85
desc = ("Extract neighborhood",
"of *window* size",
"Count gene pairs within *window*",
"Find regions above *score* cutoff",
"Identify flankers",
"Annotate syntelog class"
)
for yt, t in zip((.88, .84, .64, .6, .3, .26), desc):
root.text(xt, yt, markup(t), ha="center", va="center")
root.set_xlim(0, 1)
root.set_ylim(0, 1)
root.set_axis_off()
pf = "cartoon"
image_name = pf + "." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
def plot(args):
"""
%prog plot input.bed seqid
Plot the matchings between the reconstructed pseudomolecules and the maps.
Two types of visualizations are available in one canvas:
1. Parallel axes, and matching markers are shown in connecting lines;
2. Scatter plot.
"""
from jcvi.graphics.base import plt, savefig, normalize_axes, \
set2, panel_labels
from jcvi.graphics.chromosome import Chromosome, GeneticMap, \
HorizontalChromosome
p = OptionParser(plot.__doc__)
add_allmaps_plot_options(p)
opts, args, iopts = p.set_image_options(args, figsize="10x6")
if len(args) != 2:
sys.exit(not p.print_help())
inputbed, seqid = args
pf = inputbed.rsplit(".", 1)[0]
bedfile = pf + ".lifted.bed"
agpfile = pf + ".agp"
weightsfile = opts.weightsfile
links = opts.links
function = get_function(opts.distance)
cc = Map(bedfile, function)
allseqids = cc.seqids
mapnames = cc.mapnames
weights = Weights(weightsfile, mapnames)
assert seqid in allseqids, "{0} not in {1}".format(seqid, allseqids)
s = Scaffold(seqid, cc)
mlgs = [k for k, v in s.mlg_counts.items() if v >= links]
while not mlgs:
links /= 2
logging.error("No markers to plot, --links reset to {0}".format(links))
mlgs = [k for k, v in s.mlg_counts.items() if v >= links]
mlgsizes = {}
for mlg in mlgs:
mm = cc.extract_mlg(mlg)
mlgsize = max(function(x) for x in mm)
mlgsizes[mlg] = mlgsize
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
ax1 = fig.add_axes([0, 0, .5, 1])
ax2 = fig.add_axes([.5, 0, .5, 1])
# Find the layout first
ystart, ystop = .9, .1
L = Layout(mlgsizes)
coords = L.coords
tip = .02
marker_pos = {}
# Palette
colors = dict((mapname, set2[i]) for i, mapname in enumerate(mapnames))
colors = dict((mlg, colors[mlg.split("-")[0]]) for mlg in mlgs)
rhos = {}
# Parallel coordinates
for mlg, (x, y1, y2) in coords.items():
mm = cc.extract_mlg(mlg)
markers = [(m.accn, function(m)) for m in mm] # exhaustive marker list
xy = [(m.pos, function(m)) for m in mm if m.seqid == seqid]
mx, my = zip(*xy)
rho = spearmanr(mx, my)
rhos[mlg] = rho
flip = rho < 0
g = GeneticMap(ax1, x, y1, y2, markers, tip=tip, flip=flip)
extra = -3 * tip if x < .5 else 3 * tip
ha = "right" if x < .5 else "left"
mapname = mlg.split("-")[0]
tlg = mlg.replace("_", ".") # Latex does not like underscore char
label = "{0} (w={1})".format(tlg, weights[mapname])
ax1.text(x + extra, (y1 + y2) / 2, label, color=colors[mlg],
ha=ha, va="center", rotation=90)
marker_pos.update(g.marker_pos)
agp = AGP(agpfile)
agp = [x for x in agp if x.object == seqid]
chrsize = max(x.object_end for x in agp)
# Pseudomolecules in the center
r = ystart - ystop
ratio = r / chrsize
f = lambda x: (ystart - ratio * x)
patchstart = [f(x.object_beg) for x in agp if not x.is_gap]
Chromosome(ax1, .5, ystart, ystop, width=2 * tip, patch=patchstart, lw=2)
label = "{0} ({1})".format(seqid, human_size(chrsize, precision=0))
ax1.text(.5, ystart + tip, label, ha="center")
scatter_data = defaultdict(list)
# Connecting lines
for b in s.markers:
marker_name = b.accn
if marker_name not in marker_pos:
continue
cx = .5
cy = f(b.pos)
mx = coords[b.mlg][0]
my = marker_pos[marker_name]
extra = -tip if mx < cx else tip
extra *= 1.25 # leave boundaries for aesthetic reasons
cx += extra
mx -= extra
ax1.plot((cx, mx), (cy, my), "-", color=colors[b.mlg])
scatter_data[b.mlg].append((b.pos, function(b)))
# Scatter plot, same data as parallel coordinates
xstart, xstop = sorted((ystart, ystop))
f = lambda x: (xstart + ratio * x)
pp = [x.object_beg for x in agp if not x.is_gap]
patchstart = [f(x) for x in pp]
HorizontalChromosome(ax2, xstart, xstop, ystop,
height=2 * tip, patch=patchstart, lw=2)
gap = .03
ratio = (r - gap * len(mlgs) - tip) / sum(mlgsizes.values())
tlgs = []
for mlg, mlgsize in sorted(mlgsizes.items()):
height = ratio * mlgsize
ystart -= height
xx = .5 + xstart / 2
width = r / 2
color = colors[mlg]
ax = fig.add_axes([xx, ystart, width, height])
ypos = ystart + height / 2
ystart -= gap
sd = scatter_data[mlg]
xx, yy = zip(*sd)
ax.vlines(pp, 0, mlgsize, colors="beige")
ax.plot(xx, yy, ".", color=color)
rho = rhos[mlg]
ax.text(.5, 1 - .4 * gap / height, r"$\rho$={0:.3f}".format(rho),
ha="center", va="top", transform=ax.transAxes, color="gray")
tlg = mlg.replace("_", ".")
tlgs.append((tlg, ypos, color))
ax.set_xlim(0, chrsize)
ax.set_ylim(0, mlgsize)
ax.set_xticks([])
while height / len(ax.get_yticks()) < .03 and len(ax.get_yticks()) >= 2:
ax.set_yticks(ax.get_yticks()[::2]) # Sparsify the ticks
yticklabels = [int(x) for x in ax.get_yticks()]
ax.set_yticklabels(yticklabels, family='Helvetica')
if rho < 0:
ax.invert_yaxis()
for i, (tlg, ypos, color) in enumerate(tlgs):
ha = "center"
if len(tlgs) > 4:
ha = "right" if i % 2 else "left"
root.text(.5, ypos, tlg, color=color, rotation=90,
ha=ha, va="center")
if opts.panels:
labels = ((.04, .96, 'A'), (.48, .96, 'B'))
panel_labels(root, labels)
normalize_axes((ax1, ax2, root))
image_name = seqid + "." + iopts.format
savefig(image_name, dpi=iopts.dpi, iopts=iopts)
plt.close(fig)
def multihistogram(args):
"""
%prog multihistogram *.histogram species
Plot the histogram based on a set of K-mer hisotograms. The method is based
on Star et al.'s method (Atlantic Cod genome paper).
"""
p = OptionParser(multihistogram.__doc__)
p.add_option("--kmin",
default=15,
type="int",
help="Minimum K-mer size, inclusive")
p.add_option("--kmax",
default=30,
type="int",
help="Maximum K-mer size, inclusive")
p.add_option("--vmin",
default=2,
type="int",
help="Minimum value, inclusive")
p.add_option("--vmax",
default=100,
type="int",
help="Maximum value, inclusive")
opts, args, iopts = p.set_image_options(args, figsize="10x5", dpi=300)
if len(args) < 1:
sys.exit(not p.print_help())
histfiles = args[:-1]
species = args[-1]
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
A = fig.add_axes([0.08, 0.12, 0.38, 0.76])
B = fig.add_axes([0.58, 0.12, 0.38, 0.76])
lines = []
legends = []
genomesizes = []
for histfile in histfiles:
ks = KmerSpectrum(histfile)
x, y = ks.get_xy(opts.vmin, opts.vmax)
K = get_number(op.basename(histfile).split(".")[0].split("-")[-1])
if not opts.kmin <= K <= opts.kmax:
continue
(line, ) = A.plot(x, y, "-", lw=1)
lines.append(line)
legends.append("K = {0}".format(K))
ks.analyze(K=K, method="allpaths")
genomesizes.append((K, ks.genomesize / 1e6))
leg = A.legend(lines, legends, shadow=True, fancybox=True)
leg.get_frame().set_alpha(0.5)
title = "{0} genome K-mer histogram".format(species)
A.set_title(markup(title))
xlabel, ylabel = "Coverage (X)", "Counts"
A.set_xlabel(xlabel)
A.set_ylabel(ylabel)
set_human_axis(A)
title = "{0} genome size estimate".format(species)
B.set_title(markup(title))
x, y = zip(*genomesizes)
B.plot(x, y, "ko", mfc="w")
t = np.linspace(opts.kmin - 0.5, opts.kmax + 0.5, 100)
p = np.poly1d(np.polyfit(x, y, 2))
B.plot(t, p(t), "r:")
xlabel, ylabel = "K-mer size", "Estimated genome size (Mb)"
B.set_xlabel(xlabel)
B.set_ylabel(ylabel)
set_ticklabels_helvetica(B)
labels = ((0.04, 0.96, "A"), (0.54, 0.96, "B"))
panel_labels(root, labels)
normalize_axes(root)
imagename = species + ".multiK.pdf"
savefig(imagename, dpi=iopts.dpi, iopts=iopts)
def multihistogram(args):
"""
%prog multihistogram *.histogram species
Plot the histogram based on a set of K-mer hisotograms. The method is based
on Star et al.'s method (Atlantic Cod genome paper).
"""
p = OptionParser(multihistogram.__doc__)
p.add_option("--kmin", default=15, type="int", help="Minimum K-mer size, inclusive")
p.add_option("--kmax", default=30, type="int", help="Maximum K-mer size, inclusive")
p.add_option("--vmin", default=2, type="int", help="Minimum value, inclusive")
p.add_option("--vmax", default=100, type="int", help="Maximum value, inclusive")
opts, args, iopts = p.set_image_options(args, figsize="10x5", dpi=300)
histfiles = args[:-1]
species = args[-1]
fig = plt.figure(1, (iopts.w, iopts.h))
root = fig.add_axes([0, 0, 1, 1])
A = fig.add_axes([0.08, 0.12, 0.38, 0.76])
B = fig.add_axes([0.58, 0.12, 0.38, 0.76])
lines = []
legends = []
genomesizes = []
for histfile in histfiles:
ks = KmerSpectrum(histfile)
x, y = ks.get_xy(opts.vmin, opts.vmax)
K = get_number(op.basename(histfile).split(".")[0].split("-")[-1])
if not opts.kmin <= K <= opts.kmax:
continue
line, = A.plot(x, y, "-", lw=1)
lines.append(line)
legends.append("K = {0}".format(K))
ks.analyze(K=K)
genomesizes.append((K, ks.genomesize / 1e6))
leg = A.legend(lines, legends, shadow=True, fancybox=True)
leg.get_frame().set_alpha(0.5)
title = "{0} genome K-mer histogram".format(species)
A.set_title(markup(title))
xlabel, ylabel = "Coverage (X)", "Counts"
A.set_xlabel(xlabel)
A.set_ylabel(ylabel)
set_human_axis(A)
title = "{0} genome size estimate".format(species)
B.set_title(markup(title))
x, y = zip(*genomesizes)
B.plot(x, y, "ko", mfc="w")
t = np.linspace(opts.kmin - 0.5, opts.kmax + 0.5, 100)
p = np.poly1d(np.polyfit(x, y, 2))
B.plot(t, p(t), "r:")
xlabel, ylabel = "K-mer size", "Estimated genome size (Mb)"
B.set_xlabel(xlabel)
B.set_ylabel(ylabel)
set_ticklabels_helvetica(B)
labels = ((0.04, 0.96, "A"), (0.54, 0.96, "B"))
panel_labels(root, labels)
normalize_axes(root)
imagename = species + ".multiK.pdf"
savefig(imagename, dpi=iopts.dpi, iopts=iopts)
