def geneusage_ttests(attr, type, outfile, g2n, name2obj, matched, pcutoff):
# attr: "clones"/"reads"; type=genetype: v, d, j, vj, dj
testtype = statcommon.get_test_type(g2n, matched)
f = open(outfile, 'w')
f.write(("Name\tGroup1_Group2\tt_val\tp_val\tMean1 +/- Std1\t" +
"Mean2 +/- Std2\n"))
genes = get_genes(name2obj.values(), type)
for gene in genes:
pair2tp, group2mean = statcommon.ttest_allpairs(g2n, name2obj, matched,
attr=None, func=get_geneusage,
func_args=(attr, type, gene),
testtype=testtype)
statcommon.ttest_write(f, gene, pair2tp, group2mean, pcutoff)
f.close()
def run(self):
testtype = statcommon.get_test_type(self.g2n, self.matched)
p2stat, g2mean = statcommon.ttest_allpairs(self.g2n, self.name2stat,
self.matched, self.index,
testtype=testtype)
picklefile = os.path.join(self.outdir, "%s.pickle" % self.index)
pickle.dump((p2stat, g2mean), gzip.open(picklefile, 'wb'))
if self.plotfmt:
plotfile = os.path.join(self.plotdir, self.index)
if not self.matched:
dvplot.draw_diversity_plot(self.g2n, self.name2stat,
self.index, plotfile, self.plotfmt)
else:
dvplot.draw_diversity_plot_hacktimeseries(self.g2n,
self.name2stat, self.index, plotfile, self.plotfmt)
def lendist_ttests(attr, outfile, g2n, name2obj, matched, pcutoff):
testtype = statcommon.get_test_type(g2n, matched)
f = open(outfile, 'w')
# compare median length
f.write(("#Category\tGroup1_Group2\tt_val\tp_val\tMean1 +/- Std1\t" +
"Mean2 +/- Std2\n"))
medattr = "median_" + attr
pair2tp, group2mean = statcommon.ttest_allpairs(g2n, name2obj, matched,
medattr, testtype=testtype)
statcommon.ttest_write(f, medattr, pair2tp, group2mean)
# compare freq of each length
f.write("#Length\n")
lens = get_lens(name2obj.values())
lenattr = "len2" + attr
for l in lens:
pair2tp, group2mean = statcommon.ttest_allpairs(g2n, name2obj,
matched, attr=None,
func=statcommon.obj_dictattr_lookup,
func_args=(lenattr, l), testtype=testtype)
statcommon.ttest_write(f, str(l), pair2tp, group2mean, pcutoff)
f.close()
