def Run(self):
#if not self.wxobj:
# # Force matplotlib to use good backend for png.
# import matplotlib.pyplot as plt
#elif "matplotlib.pyplot" not in sys.modules:
try:
import matplotlib.pyplot as plt
except:
print "Error: cannot do histograms"
self.doHistogram = False
self.transit_message("Starting resampling Method")
start_time = time.time()
if self.doHistogram:
histPath = os.path.join(
os.path.dirname(self.output.name),
transit_tools.fetch_name(self.output.name) + "_histograms")
if not os.path.isdir(histPath):
os.makedirs(histPath)
else:
histPath = ""
Kctrl = len(self.ctrldata)
Kexp = len(self.expdata)
#Get orf data
self.transit_message("Getting Data")
(data, position) = transit_tools.get_validated_data(self.ctrldata +
self.expdata,
wxobj=self.wxobj)
(K, N) = data.shape
if self.normalization != "nonorm":
self.transit_message("Normalizing using: %s" % self.normalization)
(data,
factors) = norm_tools.normalize_data(data, self.normalization,
self.ctrldata + self.expdata,
self.annotation_path)
if self.LOESS:
self.transit_message("Performing LOESS Correction")
for j in range(K):
data[j] = stat_tools.loess_correction(position, data[j])
G = tnseq_tools.Genes(self.ctrldata + self.expdata,
self.annotation_path,
ignoreCodon=self.ignoreCodon,
nterm=self.NTerminus,
cterm=self.CTerminus,
data=data,
position=position)
#G = tnseq_tools.Genes(self.ctrldata+self.expdata, self.annotation_path, norm=self.normalization, ignoreCodon=self.ignoreCodon, nterm=self.NTerminus, cterm=self.CTerminus)
#Resampling
data = []
N = len(G)
count = 0
self.progress_range(N)
for gene in G:
count += 1
if gene.k == 0 or gene.n == 0:
(test_obs, mean1, mean2, log2FC, pval_ltail, pval_utail,
pval_2tail, testlist, data1, data2) = (0, 0, 0, 0, 1.00, 1.00,
1.00, [], [0], [0])
else:
if not self.includeZeros:
ii = numpy.sum(gene.reads, 0) > 0
else:
ii = numpy.ones(gene.n) == 1
data1 = gene.reads[:Kctrl, ii].flatten() + self.pseudocount
data2 = gene.reads[Kctrl:, ii].flatten() + self.pseudocount
(test_obs, mean1, mean2, log2FC, pval_ltail, pval_utail,
pval_2tail, testlist) = stat_tools.resampling(
data1,
data2,
S=self.samples,
testFunc=stat_tools.F_mean_diff_flat,
adaptive=self.adaptive)
if self.doHistogram:
import matplotlib.pyplot as plt
if testlist:
n, bins, patches = plt.hist(testlist,
density=1,
facecolor='c',
alpha=0.75,
bins=100)
else:
n, bins, patches = plt.hist([0, 0],
density=1,
facecolor='c',
alpha=0.75,
bins=100)
plt.xlabel('Delta Mean')
plt.ylabel('Probability')
plt.title('%s - Histogram of Delta Mean' % gene.orf)
plt.axvline(test_obs,
color='r',
linestyle='dashed',
linewidth=3)
plt.grid(True)
genePath = os.path.join(histPath, gene.orf + ".png")
if not os.path.exists(histPath):
os.makedirs(histPath)
plt.savefig(genePath)
plt.clf()
sum1 = numpy.sum(data1)
sum2 = numpy.sum(data2)
data.append([
gene.orf, gene.name, gene.desc, gene.n, mean1, mean2, sum1,
sum2, test_obs, log2FC, pval_2tail
])
# Update progress
text = "Running Resampling Method... %5.1f%%" % (100.0 * count / N)
self.progress_update(text, count)
#
self.transit_message("") # Printing empty line to flush stdout
self.transit_message("Performing Benjamini-Hochberg Correction")
data.sort()
qval = stat_tools.BH_fdr_correction([row[-1] for row in data])
self.output.write("#Resampling\n")
if self.wxobj:
members = sorted([
attr for attr in dir(self) if not callable(getattr(self, attr))
and not attr.startswith("__")
])
memberstr = ""
for m in members:
memberstr += "%s = %s, " % (m, getattr(self, m))
self.output.write(
"#GUI with: norm=%s, samples=%s, pseudocounts=%1.2f, adaptive=%s, histogram=%s, includeZeros=%s, output=%s\n"
% (self.normalization, self.samples, self.pseudocount,
self.adaptive, self.doHistogram, self.includeZeros,
self.output.name.encode('utf-8')))
else:
self.output.write("#Console: python %s\n" % " ".join(sys.argv))
self.output.write("#Control Data: %s\n" %
(",".join(self.ctrldata).encode('utf-8')))
self.output.write("#Experimental Data: %s\n" %
(",".join(self.expdata).encode('utf-8')))
self.output.write("#Annotation path: %s\n" %
(self.annotation_path.encode('utf-8')))
self.output.write("#Time: %s\n" % (time.time() - start_time))
self.output.write("#%s\n" % "\t".join(columns))
for i, row in enumerate(data):
(orf, name, desc, n, mean1, mean2, sum1, sum2, test_obs, log2FC,
pval_2tail) = row
self.output.write(
"%s\t%s\t%s\t%d\t%1.1f\t%1.1f\t%1.2f\t%1.1f\t%1.2f\t%1.1f\t%1.5f\t%1.5f\n"
% (orf, name, desc, n, mean1, mean2, log2FC, sum1, sum2,
test_obs, pval_2tail, qval[i]))
self.output.close()
self.transit_message("Adding File: %s" % (self.output.name))
self.add_file(filetype="Resampling")
self.finish()
self.transit_message("Finished resampling Method")
def Run(self):
self.transit_message("Starting Griffin Method")
start_time = time.time()
#Get orf data
self.transit_message("Getting Data")
(data, position) = transit_tools.get_validated_data(self.ctrldata,
wxobj=self.wxobj)
(K, N) = data.shape
if self.normalization and self.normalization != "nonorm":
self.transit_message("Normalizing using: %s" % self.normalization)
(data,
factors) = norm_tools.normalize_data(data, self.normalization,
self.ctrldata,
self.annotation_path)
G = tnseq_tools.Genes(self.ctrldata,
self.annotation_path,
minread=1,
reps=self.replicates,
ignoreCodon=self.ignoreCodon,
nterm=self.NTerminus,
cterm=self.CTerminus,
data=data,
position=position)
N = len(G)
self.progress_range(N)
count = 0
pins = G.global_theta()
pnon = 1.0 - pins
results = []
for gene in G:
if gene.n == 0:
results.append([gene, 0.0, 1.000])
else:
B = 1.0 / math.log(1.0 / pnon)
u = math.log(gene.n * pins, 1.0 / pnon)
exprun = tnseq_tools.ExpectedRuns(gene.n, pnon)
pval = 1.0 - tnseq_tools.GumbelCDF(gene.r, u, B)
results.append([gene, exprun, pval])
text = "Running Griffin Method... %5.1f%%" % (100.0 * (count + 1) /
(N))
self.progress_update(text, count)
count += 1
pval = [row[-1] for row in results]
padj = stat_tools.BH_fdr_correction(pval)
for i in range(len(results)):
results[i].append(padj[i])
results.sort()
self.output.write("#Griffin\n")
if self.wxobj:
members = sorted([
attr for attr in dir(self) if not callable(getattr(self, attr))
and not attr.startswith("__")
])
memberstr = ""
for m in members:
memberstr += "%s = %s, " % (m, getattr(self, m))
self.output.write(
"#GUI with: ctrldata=%s, annotation=%s, output=%s\n" %
(",".join(self.ctrldata).encode('utf-8'),
self.annotation_path.encode('utf-8'),
self.output.name.encode('utf-8')))
else:
self.output.write("#Console: python3 %s\n" % " ".join(sys.argv))
self.output.write("#Data: %s\n" %
(",".join(self.ctrldata).encode('utf-8')))
self.output.write("#Annotation path: %s\n" %
self.annotation_path.encode('utf-8'))
self.output.write("#Time: %s\n" % (time.time() - start_time))
self.output.write("#%s\n" % "\t".join(columns))
for (gene, exprun, pval, padj) in results:
self.output.write(
"%s\t%s\t%s\t%s\t%s\t%s\t%s\t%s\t%1.1f\t%1.5f\t%1.5f\n" %
(gene.orf, gene.name, gene.desc, gene.k, gene.n, gene.r,
gene.s, gene.t, exprun, pval, padj))
self.output.close()
self.transit_message("") # Printing empty line to flush stdout
self.transit_message("Adding File: %s" % (self.output.name))
self.add_file(filetype="Griffin")
self.finish()
self.transit_message("Finished Griffin Method")
def run_resampling(self, G_ctrl, G_exp = None, doLibraryResampling = False, histPath = ""):
data = []
N = len(G_ctrl)
count = 0
self.progress_range(N)
for gene in G_ctrl:
if gene.orf not in G_exp:
if self.diffStrains:
continue
else:
self.transit_error("Error: Gene in ctrl data not present in exp data")
self.transit_error("Make sure all .wig files come from the same strain.")
return ([], [])
gene_exp = G_exp[gene.orf]
count+=1
if not self.diffStrains and gene.n != gene_exp.n:
self.transit_error("Error: No. of TA sites in Exp and Ctrl data are different")
self.transit_error("Make sure all .wig files come from the same strain.")
return ([], [])
if (gene.k == 0 and gene_exp.k == 0) or gene.n == 0 or gene_exp.n == 0:
(test_obs, mean1, mean2, log2FC, pval_ltail, pval_utail, pval_2tail, testlist, data1, data2) = (0, 0, 0, 0, 1.00, 1.00, 1.00, [], [0], [0])
else:
if not self.includeZeros:
ii_ctrl = numpy.sum(gene.reads,0) > 0
ii_exp = numpy.sum(gene_exp.reads,0) > 0
else:
ii_ctrl = numpy.ones(gene.n) == 1
ii_exp = numpy.ones(gene_exp.n) == 1
#data1 = gene.reads[:,ii_ctrl].flatten() + self.pseudocount # we used to have an option to add pseudocounts to each observation, like this
data1 = gene.reads[:,ii_ctrl].flatten()
data2 = gene_exp.reads[:,ii_exp].flatten()
if doLibraryResampling:
(test_obs, mean1, mean2, log2FC, pval_ltail, pval_utail, pval_2tail, testlist) = stat_tools.resampling(data1, data2, S=self.samples, testFunc=stat_tools.F_mean_diff_dict, permFunc=stat_tools.F_shuffle_dict_libraries, adaptive=self.adaptive, lib_str1=self.ctrl_lib_str, lib_str2=self.exp_lib_str,PC=self.pseudocount)
else:
(test_obs, mean1, mean2, log2FC, pval_ltail, pval_utail, pval_2tail, testlist) = stat_tools.resampling(data1, data2, S=self.samples, testFunc=stat_tools.F_mean_diff_flat, permFunc=stat_tools.F_shuffle_flat, adaptive=self.adaptive, lib_str1=self.ctrl_lib_str, lib_str2=self.exp_lib_str,PC=self.pseudocount)
if self.doHistogram:
import matplotlib.pyplot as plt
if testlist:
n, bins, patches = plt.hist(testlist, density=1, facecolor='c', alpha=0.75, bins=100)
else:
n, bins, patches = plt.hist([0,0], density=1, facecolor='c', alpha=0.75, bins=100)
plt.xlabel('Delta Mean')
plt.ylabel('Probability')
plt.title('%s - Histogram of Delta Mean' % gene.orf)
plt.axvline(test_obs, color='r', linestyle='dashed', linewidth=3)
plt.grid(True)
genePath = os.path.join(histPath, gene.orf +".png")
if not os.path.exists(histPath):
os.makedirs(histPath)
plt.savefig(genePath)
plt.clf()
sum1 = numpy.sum(data1)
sum2 = numpy.sum(data2)
data.append([gene.orf, gene.name, gene.desc, gene.n, mean1, mean2, sum1, sum2, test_obs, log2FC, pval_2tail])
# Update progress
text = "Running Resampling Method... %5.1f%%" % (100.0*count/N)
self.progress_update(text, count)
#
self.transit_message("") # Printing empty line to flush stdout
self.transit_message("Performing Benjamini-Hochberg Correction")
data.sort()
qval = stat_tools.BH_fdr_correction([row[-1] for row in data])
return (data, qval)
def Run(self):
self.transit_message("Starting Tn5 gaps method")
start_time = time.time()
self.transit_message("Getting data (May take a while)")
# Combine all wigs
(data, position) = transit_tools.get_validated_data(self.ctrldata,
wxobj=self.wxobj)
combined = tnseq_tools.combine_replicates(data, method=self.replicates)
combined[combined < self.minread] = 0
counts = combined
counts[counts > 0] = 1
num_sites = counts.size
genes_obj = tnseq_tools.Genes(self.ctrldata,
self.annotation_path,
ignoreCodon=self.ignoreCodon,
nterm=self.NTerminus,
cterm=self.CTerminus,
data=data,
position=position)
pins = numpy.mean(counts)
pnon = 1.0 - pins
# Calculate stats of runs
exprunmax = tnseq_tools.ExpectedRuns(num_sites, pnon)
varrun = tnseq_tools.VarR(num_sites, pnon)
stddevrun = math.sqrt(varrun)
exp_cutoff = exprunmax + 2 * stddevrun
# Get the runs
self.transit_message("Getting non-insertion runs in genome")
run_arr = tnseq_tools.runs_w_info(counts)
pos_hash = transit_tools.get_pos_hash(self.annotation_path)
# Finally, calculate the results
self.transit_message("Running Tn5 gaps method")
results_per_gene = {}
for gene in genes_obj.genes:
results_per_gene[gene.orf] = [
gene.orf, gene.name, gene.desc, gene.k, gene.n, gene.r, 0, 0, 1
]
N = len(run_arr)
count = 0
accum = 0
self.progress_range(N)
for run in run_arr:
accum += run['length']
count += 1
genes = tnseq_tools.get_genes_in_range(pos_hash, run['start'],
run['end'])
for gene_orf in genes:
gene = genes_obj[gene_orf]
inter_sz = self.intersect_size([run['start'], run['end']],
[gene.start, gene.end]) + 1
percent_overlap = self.calc_overlap([run['start'], run['end']],
[gene.start, gene.end])
run_len = run['length']
B = 1.0 / math.log(1.0 / pnon)
u = math.log(num_sites * pins, 1.0 / pnon)
pval = 1.0 - tnseq_tools.GumbelCDF(run['length'], u, B)
curr_val = results_per_gene[gene.orf]
curr_inter_sz = curr_val[6]
curr_len = curr_val[7]
if inter_sz > curr_inter_sz:
results_per_gene[gene.orf] = [
gene.orf, gene.name, gene.desc, gene.k, gene.n, gene.r,
inter_sz, run_len, pval
]
# Update Progress
text = "Running Tn5Gaps method... %1.1f%%" % (100.0 * count / N)
self.progress_update(text, count)
data = list(results_per_gene.values())
exp_run_len = float(accum) / N
min_sig_len = float('inf')
sig_genes_count = 0
pval = [row[-1] for row in data]
padj = stat_tools.BH_fdr_correction(pval)
for i in range(len(data)):
if padj[i] < 0.05:
sig_genes_count += 1
min_sig_len = min(min_sig_len, data[i][-2])
data[i].append(padj[i])
data[i].append('Essential' if padj[i] < 0.05 else 'Non-essential')
#(data[i][0], data[i][1], data[i][2], data[i][3], data[i][4], data[i][5], data[i][6], data[i][7], data[i][8], padj[i], 'Essential' if padj[i] < 0.05 else 'Non-essential')
data.sort(key=lambda l: l[0])
# Output results
self.output.write("#Tn5 Gaps\n")
if self.wxobj:
members = sorted([
attr for attr in dir(self) if not callable(getattr(self, attr))
and not attr.startswith("__")
])
memberstr = ""
for m in members:
memberstr += "%s = %s, " % (m, getattr(self, m))
self.output.write(
"#GUI with: ctrldata=%s, annotation=%s, output=%s\n" %
(",".join(self.ctrldata).encode('utf-8'),
self.annotation_path.encode('utf-8'),
self.output.name.encode('utf-8')))
else:
self.output.write("#Console: python %s\n" % " ".join(sys.argv))
self.output.write("#Data: %s\n" %
(",".join(self.ctrldata).encode('utf-8')))
self.output.write("#Annotation path: %s\n" %
self.annotation_path.encode('utf-8'))
self.output.write("#Time: %s\n" % (time.time() - start_time))
self.output.write("#Essential gene count: %d\n" % (sig_genes_count))
self.output.write("#Minimum reads: %d\n" % (self.minread))
self.output.write("#Replicate combination method: %s\n" %
(self.replicates))
self.output.write("#Minimum significant run length: %d\n" %
(min_sig_len))
self.output.write("#Expected run length: %1.5f\n" % (exp_run_len))
self.output.write("#Expected max run length: %s\n" % (exprunmax))
self.output.write("#%s\n" % "\t".join(columns))
#self.output.write("#Orf\tName\tDesc\tk\tn\tr\tovr\tlenovr\tpval\tpadj\tcall\n")
for res in data:
self.output.write(
"%s\t%s\t%s\t%s\t%s\t%s\t%d\t%d\t%1.5f\t%1.5f\t%s\n" %
(res[0], res[1], res[2], res[3], res[4], res[5], res[6],
res[7], res[8], res[9], res[10]))
self.output.close()
self.transit_message("") # Printing empty line to flush stdout
self.transit_message("Adding File: %s" % (self.output.name))
self.add_file(filetype="Tn5 Gaps")
self.finish()
self.transit_message("Finished Tn5Gaps Method")
def Run(self):
self.transit_message("Starting rankproduct Method")
start_time = time.time()
Kctrl = len(self.ctrldata)
Kexp = len(self.expdata)
#Get orf data
self.transit_message("Getting Data")
(data, position) = transit_tools.get_validated_data(self.ctrldata +
self.expdata,
wxobj=self.wxobj)
if self.normalization != "none":
self.transit_message("Normalizing using: %s" % self.normalization)
(data,
factors) = norm_tools.normalize_data(data, self.normalization,
self.ctrldata + self.expdata,
self.annotation_path)
Gctrl = tnseq_tools.Genes(self.ctrldata + self.expdata,
self.annotation_path,
ignoreCodon=self.ignoreCodon,
nterm=self.NTerminus,
cterm=self.CTerminus,
data=data[:Kctrl, :],
position=position)
Gexp = tnseq_tools.Genes(self.ctrldata + self.expdata,
self.annotation_path,
ignoreCodon=self.ignoreCodon,
nterm=self.NTerminus,
cterm=self.CTerminus,
data=data[Kctrl:, :],
position=position)
Ngenes = len(Gctrl)
# Get the average counts for all the genes, in each replicate
meanCtrl = numpy.zeros((Kctrl, Ngenes))
meanExp = numpy.zeros((Kexp, Ngenes))
for i in range(Ngenes):
if numpy.any(Gctrl[i].reads):
meanCtrl[:, i] = numpy.mean(Gctrl[i].reads, 1)
else:
meanCtrl[:, i] = numpy.zeros(Kctrl)
#
if numpy.any(Gexp[i].reads):
meanExp[:, i] = numpy.mean(Gexp[i].reads, 1)
else:
meanExp[:, i] = numpy.zeros(Kexp)
# Calculate a logFC2 between Experimental and Control
# Then calculates it's rank, and observed rankProduct
logFC2 = numpy.log2((meanExp + 0.0001) / (meanCtrl + 0.0001))
rank = numpy.array([scipy.stats.rankdata(Lvec) for Lvec in logFC2])
obsRP = numpy.power(numpy.prod(rank, 0), 1.0 / Kctrl)
permutations = numpy.zeros((self.samples, Ngenes))
tempranks = scipy.array(
[numpy.arange(1, Ngenes + 1) for rep in range(Kctrl)])
for s in range(self.samples):
rankperm = numpy.array(
[numpy.random.permutation(tr) for tr in tempranks])
permutations[s] = numpy.power(numpy.prod(rankperm, 0), 1.0 / Kctrl)
rankRP = numpy.argsort(obsRP) + 1
#rankproduct
data = []
count = 0
self.progress_range(Ngenes)
for i, gene in enumerate(Gctrl):
count += 1
meanctrl = numpy.mean(Gctrl[i].reads)
meanexp = numpy.mean(Gexp[i].reads)
log2fc = numpy.log2((meanexp + 0.0001) / (meanctrl + 0.0001))
countbetter = numpy.sum(permutations <= obsRP[i])
pval = countbetter / float(self.samples * Ngenes)
e_val = countbetter / float(self.samples)
q_paper = e_val / float(rankRP[i])
data.append([
gene.orf, gene.name, gene.desc, gene.n, meanctrl, meanexp,
log2fc, obsRP[i], e_val, q_paper, pval
])
# Update Progress
text = "Running rankproduct Method... %5.1f%%" % (100.0 * count /
Ngenes)
self.progress_update(text, count)
#
self.transit_message("") # Printing empty line to flush stdout
self.transit_message("Performing Benjamini-Hochberg Correction")
data.sort()
q_bh = stat_tools.BH_fdr_correction([row[-1] for row in data])
self.output.write("#RankProduct\n")
if self.wxobj:
members = sorted([
attr for attr in dir(self) if not callable(getattr(self, attr))
and not attr.startswith("__")
])
memberstr = ""
for m in members:
memberstr += "%s = %s, " % (m, getattr(self, m))
self.output.write(
"#GUI with: ctrldata=%s, annotation=%s, output=%s\n" %
(",".join(self.ctrldata).encode('utf-8'),
self.annotation_path.encode('utf-8'),
self.output.name.encode('utf-8')))
else:
self.output.write("#Console: python %s\n" % " ".join(sys.argv))
self.output.write("#Data: %s\n" %
(",".join(self.ctrldata).encode('utf-8')))
self.output.write("#Annotation path: %s\n" %
self.annotation_path.encode('utf-8'))
self.output.write("#Time: %s\n" % (time.time() - start_time))
self.output.write("#%s\n" % (columns))
for i, row in enumerate(data):
(orf, name, desc, n, mean1, mean2, log2FCgene, obsRPgene, e_val,
q_paper, pval) = row
self.output.write(
"%s\t%s\t%s\t%d\t%1.1f\t%1.1f\t%1.2f\t%1.8f\t%1.1f\t%1.8f\n" %
(orf, name, desc, n, mean1, mean2, log2FCgene, obsRPgene,
e_val, q_paper))
self.output.close()
self.transit_message("Adding File: %s" % (self.output.name))
self.add_file(filetype="RankProduct")
self.finish()
self.transit_message("Finished rankproduct Method")
def Run(self):
self.transit_message("Starting Mann-Whitney U-test Method")
start_time = time.time()
Kctrl = len(self.ctrldata)
Kexp = len(self.expdata)
#Get orf data
self.transit_message("Getting Data")
(data, position) = transit_tools.get_validated_data(self.ctrldata+self.expdata, wxobj=self.wxobj)
(K,N) = data.shape
if self.normalization != "nonorm":
self.transit_message("Normalizing using: %s" % self.normalization)
(data, factors) = norm_tools.normalize_data(data, self.normalization, self.ctrldata+self.expdata, self.annotation_path)
if self.LOESS:
self.transit_message("Performing LOESS Correction")
for j in range(K):
data[j] = stat_tools.loess_correction(position, data[j])
G = tnseq_tools.Genes(self.ctrldata + self.expdata, self.annotation_path, ignoreCodon=self.ignoreCodon, nterm=self.NTerminus, cterm=self.CTerminus, data=data, position=position)
#u-test
data = []
N = len(G)
count = 0
self.progress_range(N)
for gene in G:
count+=1
if gene.k == 0 or gene.n == 0:
(test_obs, mean1, mean2, log2FC, u_stat, pval_2tail) = (0, 0, 0, 0, 0.0, 1.00)
else:
if not self.includeZeros:
ii = numpy.sum(gene.reads,0) > 0
else:
ii = numpy.ones(gene.n) == 1
data1 = gene.reads[:Kctrl,ii].flatten()
data2 = gene.reads[Kctrl:,ii].flatten()
try:
u_stat, pval_2tail = scipy.stats.mannwhitneyu(data1, data2,
alternative="two-sided")
except ValueError as e:
u_stat, pval_2tail = 0.0, 1.00
n1 = len(data1)
n2 = len(data2)
mean1 = 0
if n1 > 0:
mean1 = numpy.mean(data1)
mean2 = 0
if n2 > 0:
mean2 = numpy.mean(data2)
try:
# Only adjust log2FC if one of the means is zero
if mean1 > 0 and mean2 > 0:
log2FC = math.log((mean2)/(mean1),2)
else:
log2FC = math.log((mean2+1.0)/(mean1+1.0),2)
except:
log2FC = 0.0
#["Orf","Name","Desc","Sites","Mean Ctrl","Mean Exp","log2FC", "U-Statistic","p-value","Adj. p-value"]
data.append([gene.orf, gene.name, gene.desc, gene.n, mean1, mean2, log2FC, u_stat, pval_2tail])
# Update Progress
text = "Running Mann-Whitney U-test Method... %1.1f%%" % (100.0*count/N)
self.progress_update(text, count)
#
self.transit_message("") # Printing empty line to flush stdout
self.transit_message("Performing Benjamini-Hochberg Correction")
data.sort()
qval = stat_tools.BH_fdr_correction([row[-1] for row in data])
self.output.write("#utest\n")
if self.wxobj:
members = sorted([attr for attr in dir(self) if not callable(getattr(self,attr)) and not attr.startswith("__")])
memberstr = ""
for m in members:
memberstr += "%s = %s, " % (m, getattr(self, m))
self.output.write("#GUI with: norm=%s, includeZeros=%s, output=%s\n" % (self.normalization, self.includeZeros, self.output.name.encode('utf-8')))
else:
self.output.write("#Console: python %s\n" % " ".join(sys.argv))
self.output.write("#Control Data: %s\n" % (",".join(self.ctrldata).encode('utf-8')))
self.output.write("#Experimental Data: %s\n" % (",".join(self.expdata).encode('utf-8')))
self.output.write("#Annotation path: %s\n" % (self.annotation_path.encode('utf-8')))
self.output.write("#Time: %s\n" % (time.time() - start_time))
self.output.write("#%s\n" % "\t".join(columns))
for i,row in enumerate(data):
(orf, name, desc, n, mean1, mean2, log2FC, u_stat, pval_2tail) = row
self.output.write("%s\t%s\t%s\t%d\t%1.1f\t%1.1f\t%1.2f\t%1.2f\t%1.5f\t%1.5f\n" % (orf, name, desc, n, mean1, mean2, log2FC, u_stat, pval_2tail, qval[i]))
self.output.close()
self.transit_message("Adding File: %s" % (self.output.name))
self.add_file(filetype="utest")
self.finish()
self.transit_message("Finished Mann-Whitney U-test Method")
