def test_nb():
"""Test notebook code"""
godag = get_godag("go-basic.obo", optional_attrs={'relationship'})
go_leafs = set(o.item_id for o in godag.values() if not o.children)
virion = 'GO:0019012'
gosubdag_r0 = GoSubDag(go_leafs, godag)
nt_virion = gosubdag_r0.go2nt[virion]
print(nt_virion)
print('r0 THE VALUE OF dcnt IS: {dcnt}'.format(dcnt=nt_virion.dcnt))
gosubdag_r1 = GoSubDag(go_leafs, godag, relationships=True)
nt_virion = gosubdag_r1.go2nt[virion]
print(nt_virion)
print('r1 THE VALUE OF dcnt IS: {dcnt}'.format(dcnt=nt_virion.dcnt))
gosubdag_partof = GoSubDag(go_leafs, godag, relationships={'part_of'})
nt_virion = gosubdag_partof.go2nt[virion]
print(nt_virion)
print('THE VALUE OF dcnt IS: {dcnt}'.format(dcnt=nt_virion.dcnt))
virion_descendants = gosubdag_partof.rcntobj.go2descendants[virion]
print('{N} descendants of virion were found'.format(
N=len(virion_descendants)))
# Limit plot of descendants to get a smaller plot
virion_capsid_fiber = {'GO:0098033', 'GO:0098032'}
gosubdag_partof.prt_goids(virion_capsid_fiber,
'{NS} {GO} dcnt({dcnt}) D-{depth:02} {GO_name}')
# Limit plot size by choosing just two virion descendants
# Get a subset containing only a couple virion descendants and their ancestors
pltdag = GoSubDag(virion_capsid_fiber, godag, relationships={'part_of'})
pltobj = GoSubDagPlot(pltdag)
pltobj.plt_dag('virion_capsid_fiber.png')
def plt_goea_results(fout_img, goea_results, **kws):
"""Plot a single page."""
go_sources = [rec.GO for rec in goea_results]
go2obj = {rec.GO: rec.goterm for rec in goea_results}
gosubdag = GoSubDag(go_sources, go2obj, rcntobj=True)
godagplot = GoSubDagPlot(gosubdag, goea_results=goea_results, **kws)
godagplot.plt_dag(fout_img)
def plt_goids(gosubdag, fout_img, goids, **kws_plt):
"""Plot GO IDs in a DAG (Directed Acyclic Graph)."""
gosubdag_plt = GoSubDag(goids,
gosubdag.go2obj,
rcntobj=gosubdag.rcntobj,
**kws_plt)
godagplot = GoSubDagPlot(gosubdag_plt, **kws_plt)
godagplot.plt_dag(fout_img)
return godagplot
def _plot_grouped_gos(self, fout_img, pltgosusr, kws_plt, kws_dag):
gosubdag_plt = GoSubDag(pltgosusr,
self.grprobj.gosubdag.get_go2obj(pltgosusr),
self.grprobj.gosubdag.relationships,
rcntobj=self.grprobj.gosubdag.rcntobj,
go2nt=self.grprobj.gosubdag.go2nt,
**kws_dag)
godagplot = GoSubDagPlot(gosubdag_plt, **kws_plt)
godagplot.plt_dag(fout_img)
def plot_all(self, goids, name, prt=sys.stdout):
"""Create plots with various numbers of relationships."""
prt.write("\nCreate GoSubDag not loading any relationship")
gosubdag_orig = GoSubDag(goids, self.go2obj, relationships=False, prt=prt)
gosubdag_orig.prt_goids(gosubdag_orig.go2obj, prt=prt)
prt.write("{N} GO IDS".format(N=len(gosubdag_orig.go2obj)))
gopltdag = GoSubDagPlot(gosubdag_orig, mark_alt_id=True)
gopltdag.plt_dag(os.path.join(REPO, "a_relationship_{NAME}_r0.png".format(NAME=name)))
# goids.update(['GO:0007507'], ['GO:0072359'])
prt.write("\nCreate GoSubDag while loading only the 'part_of' relationship")
gosubdag = GoSubDag(goids, self.go2obj, relationships=['part_of'], prt=prt)
gosubdag.prt_goids(gosubdag.go2obj, prt=prt)
prt.write("{N} GO IDS".format(N=len(gosubdag.go2obj)))
gopltdag = GoSubDagPlot(gosubdag, mark_alt_id=True)
prt.write("GO SOURCES:")
gosubdag.prt_goids(gosubdag.go_sources, prt=prt)
gopltdag.plt_dag(os.path.join(REPO, "a_relationship_{NAME}_partof.png".format(NAME=name)))
prt.write("\nCreate GoSubDag while loading all relationships")
gosubdag = GoSubDag(goids, self.go2obj, relationships=True, prt=prt)
prt.write("ALL {N} GO IDS:".format(N=len(gosubdag.go2obj)))
gosubdag.prt_goids(gosubdag.go2obj, prt=prt)
prt.write("2 GO SOURCES:")
gosubdag.prt_goids(gosubdag.go_sources, prt=prt)
goids_new = set(gosubdag.go2obj).difference(set(gosubdag_orig.go2obj))
go2color = {go:'#d5ffff' for go in goids_new}
prt.write("{N} NEW GO IDS:".format(N=len(goids_new)))
gosubdag.prt_goids(goids_new, prt=prt)
prt.write("{N} GO IDS".format(N=len(gosubdag.go2obj)))
gopltdag = GoSubDagPlot(gosubdag, mark_alt_id=True, go2color=go2color)
gopltdag.plt_dag(os.path.join(REPO, "a_relationship_{NAME}_r1.png".format(NAME=name)))
def _plt(self, goid, exp_goids, act_goids, diff_exp, diff_act):
"""Plot GO IDs, colored by differences in expected and actual"""
fout_png = '{NAME}_{GO}.png'.format(NAME=self.name,
GO=goid.replace(':', ''))
go_sources = set.union(exp_goids, act_goids, {goid})
gosubdag = GoSubDag(go_sources, self.godag, relationships=True)
go2color = {goid: '#c8ffb0'} # xkcd light light green
for go_diff in diff_exp:
go2color[go_diff] = '#cafffb' # xkcd light light blue
for go_diff in diff_act:
go2color[go_diff] = '#ffd1df' # xkcd light pink
goploter = GoSubDagPlot(gosubdag, go2color=go2color)
goploter.plt_dag(fout_png)
def plt_goids(self, fout_img, go_sources):
"""Plot GO IDs."""
# % src/bin/go_plot.py GOs --obo=../goatools/data/i86.obo --outfile=t00.jpg --mark_alt_id
gosubdag = GoSubDag(go_sources, self.gosubdag_all.go2obj, prt=self.prt,
# rcntobj=False,
rcntobj=self.gosubdag_all.rcntobj,
go2nt=self.gosubdag_all.go2nt)
prtfmt = gosubdag.prt_attr['fmta']
goids_plt = GoSubDagPlot(gosubdag).get_goids_plt()
self.prt.write("\n{N} GO IDs\n".format(N=len(goids_plt)))
gosubdag.prt_goids(goids_plt, prtfmt=prtfmt, prt=self.prt)
objplt = GoSubDagPlot(gosubdag, mark_alt_id=True)
objplt.plt_dag(os.path.join(self.cwd, fout_img))
def plt_goids(self, fout_img, go_sources):
"""Plot GO IDs."""
# % src/bin/go_plot.py GOs --obo=../goatools/data/i86.obo --outfile=t00.jpg --mark_alt_id
gosubdag = GoSubDag(
go_sources,
self.gosubdag_all.go2obj,
prt=self.prt,
# rcntobj=False,
rcntobj=self.gosubdag_all.rcntobj,
go2nt=self.gosubdag_all.go2nt)
prtfmt = gosubdag.prt_attr['fmta']
goids_plt = GoSubDagPlot(gosubdag).get_goids_plt()
self.prt.write("\n{N} GO IDs\n".format(N=len(goids_plt)))
gosubdag.prt_goids(goids_plt, prtfmt=prtfmt, prt=self.prt)
objplt = GoSubDagPlot(gosubdag, mark_alt_id=True)
objplt.plt_dag(os.path.join(self.cwd, fout_img))
def _get_objpltg(self, goids, go2color_usr, **kws):
"""Plot grouped GO IDs."""
#### fout_img = self.get_outfile(kws['outfile'], goids)
sections = read_sections(kws['sections'], exclude_ungrouped=True)
# print ("KWWSSSSSSSS", kws)
# kws_plt = {k:v for k, v in kws.items if k in self.kws_plt}
grprobj_cur = self._get_grprobj(goids, sections)
# GO: purple=hdr-only, green=hdr&usr, yellow=usr-only
# BORDER: Black=hdr Blu=hdr&usr
grpcolor = GrouperColors(
grprobj_cur) # get_bordercolor get_go2color_users
grp_go2color = grpcolor.get_go2color_users()
grp_go2bordercolor = grpcolor.get_bordercolor()
for goid, color in go2color_usr.items():
grp_go2color[goid] = color
objcolor = Go2Color(self.gosubdag,
objgoea=None,
go2color=grp_go2color,
go2bordercolor=grp_go2bordercolor)
go2txt = GrouperPlot.get_go2txt(grprobj_cur, grp_go2color,
grp_go2bordercolor)
return GoSubDagPlot(self.gosubdag,
Go2Color=objcolor,
go2txt=go2txt,
**kws)
def show_go_dag_for_terms(terms, add_relationships=True):
if type(terms) is pd.core.series.Series:
terms = terms.tolist()
if not terms:
return
with open(os.devnull, 'w') as null, redirect_stdout(null):
obo_fname = download_and_move_go_basic_obo(prt=null)
file_gene2go = download_ncbi_associations(prt=null)
if add_relationships:
optional_attrs = ['relationship', 'def']
else:
optional_attrs = ['def']
obodag = GODag("geneinfo_cache/go-basic.obo",
optional_attrs=optional_attrs,
prt=null)
gosubdag = GoSubDag(terms, obodag, relationships=add_relationships)
GoSubDagPlot(gosubdag).plt_dag('geneinfo_cache/plot.png')
return Image('geneinfo_cache/plot.png')
def test_tcntobj_relationships(do_plt=False):
"""Test loading of relationships, like part_of, into TermCounts"""
# Filenames
fin_obo = os.path.join(REPO, "tests/data/yangRWC/fig2a.obo")
fin_anno = os.path.join(REPO, "tests/data/yangRWC/fig2a.anno")
fout_png_r0 = os.path.join(REPO, 'yang_fig2a_r0.png')
fout_png_r1 = os.path.join(REPO, 'yang_fig2a_r1.png')
relationships = {
'part_of',
}
# Load ontologies
go2obj = GODag(fin_obo, optional_attrs=['relationship'])
# Load annotations
assoc = IdToGosReader(fin_anno, godag=go2obj).get_id2gos('CC')
# Count genes annotated to GO terms w and wo/relationships
tcntobj_r0 = TermCounts(go2obj, assoc)
# relationship: G (GO:0000007) is part_of F (GO:0000006)
tcntobj_r1 = TermCounts(go2obj, assoc, relationships)
# Check results
# Adding relationships does not change the total count of genes:
assert tcntobj_r0.gocnts['GO:0005575'] == tcntobj_r1.gocnts['GO:0005575']
# Counts without relationships:
assert tcntobj_r0.gocnts['GO:0000002'] == 40 # GO Term B
assert tcntobj_r0.gocnts['GO:0000006'] == 10 # GO Term F
# Counts with relationships: F counts G's 30 genes, so does B
assert tcntobj_r1.gocnts['GO:0000002'] == 70 # GO Term B
assert tcntobj_r1.gocnts['GO:0000006'] == 40 # GO Term F
# Optionally visualize the difference between term counts w and wo/relationships
if do_plt:
go2txt_r0 = {
nt.GO: 'tcnt={}'.format(nt.tcnt)
for nt in tcntobj_r0.gosubdag.go2nt.values()
}
GoSubDagPlot(tcntobj_r0.gosubdag,
go2txt=go2txt_r0).plt_dag(fout_png_r0)
go2txt_r1 = {
nt.GO: 'tcnt={}'.format(nt.tcnt)
for nt in tcntobj_r1.gosubdag.go2nt.values()
}
GoSubDagPlot(tcntobj_r1.gosubdag,
go2txt=go2txt_r1).plt_dag(fout_png_r1)
def get_gosubdagplot(self, goids=None, **kws_usr):
"""Plot GO IDs."""
if goids is None:
goids = self.grprobj.usrgos
kws_plt, kws_dag = self._get_kws_plt(goids, **kws_usr)
gosubdag = GoSubDag(goids,
self.grprobj.gosubdag.get_go2obj(goids),
self.grprobj.gosubdag.relationships,
rcntobj=self.grprobj.gosubdag.rcntobj,
go2nt=self.grprobj.gosubdag.go2nt,
**kws_dag)
return GoSubDagPlot(gosubdag, **kws_plt)
def _get_gosubdagplotnt(self, ntplt, title, go2color, pltargs):
"""Return GoSubDagPlotNt, which contains both a GoSubDagPlot object and ntobj."""
kws_plt = pltargs.get_kws_plt()
kws_plt['id'] = '"{ID}"'.format(ID=ntplt.hdrgo)
kws_plt['title'] = "{TITLE} of {M} user GOs".format(TITLE=title,
M=ntplt.tot_usrgos)
kws_plt['go2color'] = go2color
kws_plt['go2bordercolor'] = pltargs.go2bordercolor
if ntplt.parentcnt:
kws_plt["parentcnt"] = True
gosubdagplot = GoSubDagPlot(ntplt.gosubdag, **kws_plt)
return GoSubDagPlotNt(self.grprobj, gosubdagplot, ntplt)
def _plt_gogrouped(self, goids, go2color_usr, **kws):
"""Plot grouped GO IDs."""
fout_img = self.get_outfile(kws['outfile'], goids)
sections = read_sections(kws['sections'], exclude_ungrouped=True)
print("KWWSSSSSSSS", kws)
# kws_plt = {k:v for k, v in kws.items if k in self.kws_plt}
grprobj_cur = self._get_grprobj(goids, sections)
# GO: purple=hdr-only, green=hdr&usr, yellow=usr-only
# BORDER: Black=hdr Blu=hdr&usr
grpcolor = GrouperColors(
grprobj_cur) # get_bordercolor get_go2color_users
grp_go2color = grpcolor.get_go2color_users()
grp_go2bordercolor = grpcolor.get_bordercolor()
for goid, color in go2color_usr.items():
grp_go2color[goid] = color
objcolor = Go2Color(self.gosubdag,
objgoea=None,
go2color=grp_go2color,
go2bordercolor=grp_go2bordercolor)
go2txt = GrouperPlot.get_go2txt(grprobj_cur, grp_go2color,
grp_go2bordercolor)
objplt = GoSubDagPlot(self.gosubdag,
Go2Color=objcolor,
go2txt=go2txt,
**kws)
objplt.prt_goids(sys.stdout)
objplt.plt_dag(fout_img)
sys.stdout.write("{N:>6} sections read\n".format(
N="NO" if sections is None else len(sections)))
return fout_img
def _plt_gosubdag(self, goids, go2color, **kws):
"""Plot GO IDs."""
fout_img = self.get_outfile(kws['outfile'], goids)
objcolor = Go2Color(self.gosubdag, objgoea=None, go2color=go2color)
objplt = GoSubDagPlot(self.gosubdag, Go2Color=objcolor, **kws)
objplt.prt_goids(sys.stdout)
objplt.plt_dag(fout_img)
return fout_img
def _plt_gosubdag(self, goids, go2color, **kws):
"""Plot GO IDs."""
print("PLOTTING KWS", kws)
fout_img = self.get_outfile(kws['outfile'], goids, 'relationship' in kws)
objcolor = Go2Color(self.gosubdag, objgoea=None, go2color=go2color)
print("kws:")
print(kws)
objplt = GoSubDagPlot(self.gosubdag, Go2Color=objcolor, **kws)
objplt.prt_goids(sys.stdout)
objplt.plt_dag(fout_img)
return fout_img
def plt_goids(self, fout_img, go_sources):
"""Plot GO IDs."""
# % src/bin/go_plot.py GOs --obo=../goatools/data/i86.obo --outfile=t00.jpg --mark_alt_id
gosubdag = GoSubDag(go_sources, self.go2obj_all)
objplt = GoSubDagPlot(gosubdag, mark_alt_id=True)
objplt.plt_dag(os.path.join(self.cwd, fout_img))
def _do_plt(tcntobj, godag):
"""Plot the test GO-DAG"""
gosubdag = GoSubDag(tcntobj.go2obj.keys(), godag, tcntobj=tcntobj)
GoSubDagPlot(gosubdag).plt_dag('i148b.png')
def _get_goobjplt(gosubdag):
"""STEP 3) Get a plotting object."""
go_sources = set(["GO:0036476", "GO:0007516"])
gopltdag = GoSubDag(go_sources, gosubdag.go2obj)
return GoSubDagPlot(gopltdag)
def plt(self, fout_png, goids, relationships, go2txt):
"""Plot GO terms above resercher GO terms"""
gosubdag = GoSubDag(goids, self.godag, relationships)
GoSubDagPlot(gosubdag, go2txt=go2txt).plt_dag(fout_png)
def plot_gos(fout_img, goids, go2obj, **kws):
"""Given GO ids and the obo_dag, create a plot of paths from GO ids."""
gosubdag = GoSubDag(goids, go2obj, rcntobj=True)
godagplot = GoSubDagPlot(gosubdag, **kws)
godagplot.plt_dag(fout_img)
def _get_objplt(self, go2color, **kws):
"""Plot GO IDs."""
#### fout_img = self.get_outfile(kws['outfile'], goids)
objcolor = Go2Color(self.gosubdag, objgoea=None, go2color=go2color)
return GoSubDagPlot(self.gosubdag, Go2Color=objcolor, **kws)
def plt_goids(self, fout_img, go_sources):
"""Plot GO IDs."""
# % src/bin/go_plot.py GOs --obo=../goatools/data/i86.obo --outfile=t00.jpg --mark_alt_id
gosubdag = GoSubDag(go_sources, self.go2obj_all)
objplt = GoSubDagPlot(gosubdag, mark_alt_id=True)
objplt.plt_dag(os.path.join(self.cwd, fout_img))
def plot_all(self, goids, name, prt=sys.stdout):
"""Create plots with various numbers of relationships."""
prt.write("\nCreate GoSubDag not loading any relationship")
gosubdag_orig = GoSubDag(goids,
self.go2obj,
relationships=False,
prt=prt)
gosubdag_orig.prt_goids(gosubdag_orig.go2obj, prt=prt)
prt.write("{N} GO IDS".format(N=len(gosubdag_orig.go2obj)))
gopltdag = GoSubDagPlot(gosubdag_orig, mark_alt_id=True)
gopltdag.plt_dag(
os.path.join(REPO,
"a_relationship_{NAME}_r0.png".format(NAME=name)))
# goids.update(['GO:0007507'], ['GO:0072359'])
prt.write(
"\nCreate GoSubDag while loading only the 'part_of' relationship")
gosubdag = GoSubDag(goids,
self.go2obj,
relationships=['part_of'],
prt=prt)
gosubdag.prt_goids(gosubdag.go2obj, prt=prt)
prt.write("{N} GO IDS".format(N=len(gosubdag.go2obj)))
gopltdag = GoSubDagPlot(gosubdag, mark_alt_id=True)
prt.write("GO SOURCES:")
gosubdag.prt_goids(gosubdag.go_sources, prt=prt)
gopltdag.plt_dag(
os.path.join(REPO,
"a_relationship_{NAME}_partof.png".format(NAME=name)))
prt.write("\nCreate GoSubDag while loading all relationships")
gosubdag = GoSubDag(goids, self.go2obj, relationships=True, prt=prt)
prt.write("ALL {N} GO IDS:".format(N=len(gosubdag.go2obj)))
gosubdag.prt_goids(gosubdag.go2obj, prt=prt)
prt.write("2 GO SOURCES:")
gosubdag.prt_goids(gosubdag.go_sources, prt=prt)
goids_new = set(gosubdag.go2obj).difference(set(gosubdag_orig.go2obj))
go2color = {go: '#d5ffff' for go in goids_new}
prt.write("{N} NEW GO IDS:".format(N=len(goids_new)))
gosubdag.prt_goids(goids_new, prt=prt)
prt.write("{N} GO IDS".format(N=len(gosubdag.go2obj)))
gopltdag = GoSubDagPlot(gosubdag, mark_alt_id=True, go2color=go2color)
gopltdag.plt_dag(
os.path.join(REPO,
"a_relationship_{NAME}_r1.png".format(NAME=name)))