def Run(self):
self.transit_message("Starting Normalization")
start_time = time.time()
infile = self.infile
outputPath = self.outfile # output file exists, should I require -overwrite flag?
# determine ref genome from first; assume they are all the same; assume wigs have 2 header lines
line2 = "variableStep chrom=" # unknown
for line in open(infile):
if line.startswith("variableStep"): line2 = line.rstrip(); break
if self.combined_wig==True: (sites,data,files) = tnseq_tools.read_combined_wig(self.ctrldata[0])
else: (data, sites) = tnseq_tools.get_data(self.ctrldata)
(data,factors) = norm_tools.normalize_data(data,self.normalization)
print "writing",outputPath
file = open(outputPath,"w")
file.write("# %s normalization of %s\n" % (self.normalization,infile))
if self.combined_wig==True:
for f in files: file.write("#File: %s\n" % f)
for i in range(len(sites)): file.write('\t'.join([str(sites[i])]+["%0.1f" % x for x in list(data[...,i])])+"\n")
else:
file.write(line2+"\n")
for j in range(len(sites)):
file.write("%s %s\n" % (sites[j],int(data[0,j])))
file.close()
self.finish()
self.transit_message("Finished Normalization")
def Run(self):
self.transit_message("Starting Normalization")
start_time = time.time()
infile = self.infile
outputPath = self.outfile # output file exists, should I require -overwrite flag?
# determine ref genome from first; assume they are all the same; assume wigs have 2 header lines
line2 = "variableStep chrom=" # unknown
for line in open(infile):
if line.startswith("variableStep"): line2 = line.rstrip(); break
if self.combined_wig==True: (sites,data,files) = tnseq_tools.read_combined_wig(self.ctrldata[0])
else: (data, sites) = tnseq_tools.get_data(self.ctrldata)
(data,factors) = norm_tools.normalize_data(data,self.normalization)
print "writing",outputPath
file = open(outputPath,"w")
file.write("# %s normalization of %s\n" % (self.normalization,infile))
if self.combined_wig==True:
for f in files: file.write("#File: %s\n" % f)
for i in range(len(sites)): file.write('\t'.join([str(sites[i])]+["%0.1f" % x for x in list(data[...,i])])+"\n")
else:
file.write(line2+"\n")
for j in range(len(sites)):
file.write("%s %s\n" % (sites[j],int(data[0,j])))
file.close()
self.finish()
self.transit_message("Finished Normalization")
def Run(self):
self.transit_message("Starting Anova analysis")
start_time = time.time()
self.transit_message("Getting Data")
(sites, data, filenamesInCombWig) = tnseq_tools.read_combined_wig(self.combined_wig)
self.transit_message("Normalizing using: %s" % self.normalization)
(data, factors) = norm_tools.normalize_data(data, self.normalization)
conditionsByFile, _, _, orderingMetadata = tnseq_tools.read_samples_metadata(self.metadata)
conditions = self.wigs_to_conditions(
conditionsByFile,
filenamesInCombWig)
conditionsList = self.select_conditions(conditions,self.included_conditions,self.ignored_conditions,orderingMetadata)
data, conditions, _, _ = self.filter_wigs_by_conditions2(data, conditions, conditionsList)
genes = tnseq_tools.read_genes(self.annotation_path)
TASiteindexMap = {TA: i for i, TA in enumerate(sites)}
RvSiteindexesMap = tnseq_tools.rv_siteindexes_map(genes, TASiteindexMap, nterm=self.NTerminus, cterm=self.CTerminus)
MeansByRv = self.means_by_rv(data, RvSiteindexesMap, genes, conditions)
self.transit_message("Running Anova")
pvals,qvals,run_status = self.run_anova(data, genes, MeansByRv, RvSiteindexesMap, conditions)
self.transit_message("Adding File: %s" % (self.output))
file = open(self.output,"w")
heads = ("Rv Gene TAs".split() +
["Mean_%s" % x for x in conditionsList] +
["LFC_%s" % x for x in conditionsList] +
"pval padj".split() + ["status"])
file.write("#Console: python3 %s\n" % " ".join(sys.argv))
file.write("#parameters: normalization=%s, trimming=%s/%s%% (N/C), pseudocounts=%s\n" % (self.normalization,self.NTerminus,self.CTerminus,self.PC))
file.write('#'+'\t'.join(heads)+EOL)
for gene in genes:
Rv = gene["rv"]
if Rv in MeansByRv:
means = [MeansByRv[Rv][c] for c in conditionsList]
LFCs = self.calcLFCs(means,self.PC)
vals = ([Rv, gene["gene"], str(len(RvSiteindexesMap[Rv]))] +
["%0.2f" % x for x in means] +
["%0.3f" % x for x in LFCs] +
["%f" % x for x in [pvals[Rv], qvals[Rv]]] + [run_status[Rv]])
file.write('\t'.join(vals)+EOL)
file.close()
self.transit_message("Finished Anova analysis")
self.transit_message("Time: %0.1fs\n" % (time.time() - start_time))
def Run(self):
self.transit_message("Starting TnseqStats")
start_time = time.time()
datasets = self.wigs
if self.combined_wig == None:
(data, sites) = tnseq_tools.get_data(self.wigs)
else:
(sites, data,
datasets) = tnseq_tools.read_combined_wig(self.combined_wig)
# write table of stats (saturation,NZmean)
file = sys.stdout
if self.outfile != None: file = open(self.outfile, "w")
PTI = True
if PTI == True:
file.write(
"dataset\tdensity\tmean_ct\tNZmean\tNZmedian\tmax_ct\ttotal_cts\tskewness\tkurtosis\tpickands_tail_index\n"
)
else:
file.write(
"dataset\tdensity\tmean_ct\tNZmean\tNZmedian\tmax_ct\ttotal_cts\tskewness\tkurtosis\n"
)
for i in range(data.shape[0]):
density, meanrd, nzmeanrd, nzmedianrd, maxrd, totalrd, skew, kurtosis = tnseq_tools.get_data_stats(
data[i, :])
nzmedianrd = int(nzmedianrd) if numpy.isnan(
nzmedianrd) == False else 0
pti = self.pickands_tail_index(data[i, :])
vals = [
datasets[i],
"%0.3f" % density,
"%0.1f" % meanrd,
"%0.1f" % nzmeanrd,
"%d" % nzmedianrd, maxrd,
int(totalrd),
"%0.1f" % skew,
"%0.1f" % kurtosis
]
if PTI == True: vals.append("%0.3f" % pti)
file.write('\t'.join([str(x) for x in vals]) + '\n')
if self.outfile != None: file.close()
self.finish()
self.transit_message("Finished TnseqStats")
def Run(self):
self.transit_message("Starting Anova analysis")
start_time = time.time()
self.transit_message("Getting Data")
(sites, data,
filenamesInCombWig) = tnseq_tools.read_combined_wig(self.combined_wig)
self.transit_message("Normalizing using: %s" % self.normalization)
(data, factors) = norm_tools.normalize_data(data, self.normalization)
conditions = self.wigs_to_conditions(
self.read_samples_metadata(self.metadata), filenamesInCombWig)
data, conditions = self.filter_by_conditions_blacklist(
data, conditions, self.ignored_conditions)
genes = tnseq_tools.read_genes(self.annotation_path)
TASiteindexMap = {TA: i for i, TA in enumerate(sites)}
RvSiteindexesMap = tnseq_tools.rv_siteindexes_map(
genes, TASiteindexMap)
MeansByRv = self.means_by_rv(data, RvSiteindexesMap, genes, conditions)
self.transit_message("Running Anova")
pvals, qvals = self.run_anova(data, genes, MeansByRv, RvSiteindexesMap,
conditions)
self.transit_message("Adding File: %s" % (self.output))
file = open(self.output, "w")
conditionsList = list(set(conditions))
vals = "Rv Gene TAs".split() + conditionsList + "pval padj".split()
file.write('\t'.join(vals) + EOL)
for gene in genes:
Rv = gene["rv"]
if Rv in MeansByRv:
vals = ([Rv, gene["gene"],
str(len(RvSiteindexesMap[Rv]))] +
["%0.1f" % MeansByRv[Rv][c] for c in conditionsList] +
["%f" % x for x in [pvals[Rv], qvals[Rv]]])
file.write('\t'.join(vals) + EOL)
file.close()
self.transit_message("Finished Anova analysis")
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()
histPath = ""
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)
#Get orf data
self.transit_message("Getting Data")
if self.diffStrains:
self.transit_message("Multiple annotation files found")
self.transit_message("Mapping ctrl data to {0}, exp data to {1}".format(self.annotation_path, self.annotation_path_exp))
if self.combinedWigParams:
(position, data, filenamesInCombWig) = tnseq_tools.read_combined_wig(self.combinedWigParams['combined_wig'])
conditionsByFile, _, _, _ = tnseq_tools.read_samples_metadata(self.combinedWigParams['samples_metadata'])
conditions = self.wigs_to_conditions(conditionsByFile, filenamesInCombWig)
data, conditions = self.filter_wigs_by_conditions(data, conditions, self.combinedWigParams['conditions'])
data_ctrl = numpy.array([d for i, d in enumerate(data) if conditions[i].lower() == self.combinedWigParams['conditions'][0]])
data_exp = numpy.array([d for i, d in enumerate(data) if conditions[i].lower() == self.combinedWigParams['conditions'][1]])
position_ctrl, position_exp = position, position
else:
(data_ctrl, position_ctrl) = transit_tools.get_validated_data(self.ctrldata, wxobj=self.wxobj)
(data_exp, position_exp) = transit_tools.get_validated_data(self.expdata, wxobj=self.wxobj)
(K_ctrl, N_ctrl) = data_ctrl.shape
(K_exp, N_exp) = data_exp.shape
if not self.diffStrains and (N_ctrl != N_exp):
self.transit_error("Error: Ctrl and Exp wig files don't have the same number of sites.")
self.transit_error("Make sure all .wig files come from the same strain.")
return
# (data, position) = transit_tools.get_validated_data(self.ctrldata+self.expdata, wxobj=self.wxobj)
self.transit_message("Preprocessing Ctrl data...")
data_ctrl = self.preprocess_data(position_ctrl, data_ctrl)
self.transit_message("Preprocessing Exp data...")
data_exp = self.preprocess_data(position_exp, data_exp)
G_ctrl = tnseq_tools.Genes(self.ctrldata, self.annotation_path, ignoreCodon=self.ignoreCodon, nterm=self.NTerminus, cterm=self.CTerminus, data=data_ctrl, position=position_ctrl)
G_exp = tnseq_tools.Genes(self.expdata, self.annotation_path_exp, ignoreCodon=self.ignoreCodon, nterm=self.NTerminus, cterm=self.CTerminus, data=data_exp, position=position_exp)
doLibraryResampling = False
# If library string not empty
if self.ctrl_lib_str or self.exp_lib_str:
letters_ctrl = set(self.ctrl_lib_str)
letters_exp = set(self.exp_lib_str)
# Check if using exactly 1 letters; i.e. no different libraries
if len(letters_ctrl) == 1 and letters_exp==1:
pass
# If using more than one letter, then check no differences in set
else:
lib_diff = letters_ctrl ^ letters_exp
# Check that their differences
if not lib_diff:
doLibraryResampling = True
else:
transit_tools.transit_error("Error: Library Strings (Ctrl = %s, Exp = %s) do not use the same letters. Make sure every letter / library is represented in both Control and Experimental Conditions. Proceeding with resampling assuming all datasets belong to the same library." % (self.ctrl_lib_str, self.exp_lib_str))
self.ctrl_lib_str = ""
self.exp_lib_str = ""
(data, qval) = self.run_resampling(G_ctrl, G_exp, doLibraryResampling, histPath)
self.write_output(data, qval, start_time)
self.finish()
self.transit_message("Finished resampling Method")
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()
histPath = ""
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)
#Get orf data
self.transit_message("Getting Data")
if self.diffStrains:
self.transit_message("Multiple annotation files found")
self.transit_message("Mapping ctrl data to {0}, exp data to {1}".format(self.annotation_path, self.annotation_path_exp))
if self.combinedWigParams:
(position, data, filenamesInCombWig) = tnseq_tools.read_combined_wig(self.combinedWigParams['combined_wig'])
conditionsByFile, _, _, _ = tnseq_tools.read_samples_metadata(self.combinedWigParams['samples_metadata'])
conditions = self.wigs_to_conditions(conditionsByFile, filenamesInCombWig)
data, conditions = self.filter_wigs_by_conditions(data, conditions, self.combinedWigParams['conditions'])
data_ctrl = numpy.array([d for i, d in enumerate(data) if conditions[i].lower() == self.combinedWigParams['conditions'][0]])
data_exp = numpy.array([d for i, d in enumerate(data) if conditions[i].lower() == self.combinedWigParams['conditions'][1]])
position_ctrl, position_exp = position, position
else:
(data_ctrl, position_ctrl) = transit_tools.get_validated_data(self.ctrldata, wxobj=self.wxobj)
(data_exp, position_exp) = transit_tools.get_validated_data(self.expdata, wxobj=self.wxobj)
(K_ctrl, N_ctrl) = data_ctrl.shape
(K_exp, N_exp) = data_exp.shape
if not self.diffStrains and (N_ctrl != N_exp):
self.transit_error("Error: Ctrl and Exp wig files don't have the same number of sites.")
self.transit_error("Make sure all .wig files come from the same strain.")
return
# (data, position) = transit_tools.get_validated_data(self.ctrldata+self.expdata, wxobj=self.wxobj)
self.transit_message("Preprocessing Ctrl data...")
data_ctrl = self.preprocess_data(position_ctrl, data_ctrl)
self.transit_message("Preprocessing Exp data...")
data_exp = self.preprocess_data(position_exp, data_exp)
G_ctrl = tnseq_tools.Genes(self.ctrldata, self.annotation_path, ignoreCodon=self.ignoreCodon, nterm=self.NTerminus, cterm=self.CTerminus, data=data_ctrl, position=position_ctrl)
G_exp = tnseq_tools.Genes(self.expdata, self.annotation_path_exp, ignoreCodon=self.ignoreCodon, nterm=self.NTerminus, cterm=self.CTerminus, data=data_exp, position=position_exp)
doLibraryResampling = False
# If library string not empty
if self.ctrl_lib_str or self.exp_lib_str:
letters_ctrl = set(self.ctrl_lib_str)
letters_exp = set(self.exp_lib_str)
# Check if using exactly 1 letters; i.e. no different libraries
if len(letters_ctrl) == 1 and letters_exp==1:
pass
# If using more than one letter, then check no differences in set
else:
lib_diff = letters_ctrl ^ letters_exp
# Check that their differences
if not lib_diff:
doLibraryResampling = True
else:
transit_tools.transit_error("Error: Library Strings (Ctrl = %s, Exp = %s) do not use the same letters. Make sure every letter / library is represented in both Control and Experimental Conditions. Proceeding with resampling assuming all datasets belong to the same library." % (self.ctrl_lib_str, self.exp_lib_str))
self.ctrl_lib_str = ""
self.exp_lib_str = ""
(data, qval) = self.run_resampling(G_ctrl, G_exp, doLibraryResampling, histPath)
self.write_output(data, qval, start_time)
self.finish()
self.transit_message("Finished resampling Method")
def Run(self):
self.transit_message("Starting ZINB analysis")
start_time = time.time()
packnames = ("MASS", "pscl")
r_packages_needed = [
x for x in packnames if not rpackages.isinstalled(x)
]
if (len(r_packages_needed) > 0):
self.transit_error(
"Error: Following R packages are required: %(0)s. From R console, You can install them using install.packages(c(%(0)s))"
% ({
'0': '"{0}"'.format('", "'.join(r_packages_needed))
}))
sys.exit(1)
self.transit_message("Getting Data")
(sites, data,
filenamesInCombWig) = tnseq_tools.read_combined_wig(self.combined_wig)
self.transit_message("Normalizing using: %s" % self.normalization)
(data, factors) = norm_tools.normalize_data(data, self.normalization)
condition_name = self.condition
# if a covar is not found, this crashes; check for it?
conditionsByFile, covariatesByFileList, interactionsByFileList, orderingMetadata = tnseq_tools.read_samples_metadata(
self.metadata,
self.covars,
self.interactions,
condition_name=condition_name)
## [Condition] in the order of files in combined wig
conditions = self.wigs_to_conditions(conditionsByFile,
filenamesInCombWig)
## [Covariate] in the order of files in combined wig
covariates = self.wigs_to_covariates(covariatesByFileList,
filenamesInCombWig)
## [Interaction] in the order of files in combined wig
interactions = self.wigs_to_interactions(interactionsByFileList,
filenamesInCombWig)
conditionsList = self.select_conditions(conditions,
self.included_conditions,
self.ignored_conditions,
orderingMetadata)
data, conditions, covariates, interactions = self.filter_wigs_by_conditions2(
data,
conditions,
conditionsList,
covariates=covariates,
interactions=interactions)
# show the samples associated with each condition (and covariates or interactions, if defined), and count samples in each cross-product of vars
filesByCondition = self.invertDict(conditionsByFile)
samples_used = set()
for cond in conditionsList:
samples_used.update(filesByCondition[cond])
vars = [condition_name] + self.covars + self.interactions
vars2vals = {}
vars2vals[condition_name] = list(set(conditions))
for i, var in enumerate(self.covars):
vars2vals[var] = list(set(covariates[i]))
for i, var in enumerate(self.interactions):
vars2vals[var] = list(set(interactions[i]))
varsByFileList = [conditionsByFile
] + covariatesByFileList + interactionsByFileList
for i, var in enumerate(vars):
print("\nCondition/Covariate/Interaction: %s" % vars[i])
filesByVar = self.invertDict(varsByFileList[i])
for k, v in filesByVar.items():
samples = list(samples_used.intersection(set(v)))
if k in vars2vals.get(var, []):
print("%s: %s" % (k, ' '.join(samples)))
pairs = []
print("\nsamples in cross-product:")
any_empty = self.expandVar([], vars, varsByFileList, vars2vals,
set(samples_used))
if any_empty:
print(
"warning: ZINB requires samples in all combinations of conditions; the fact that one is empty could result in Model Errors"
)
genes = tnseq_tools.read_genes(self.annotation_path)
TASiteindexMap = {TA: i for i, TA in enumerate(sites)}
RvSiteindexesMap = tnseq_tools.rv_siteindexes_map(genes,
TASiteindexMap,
nterm=self.NTerminus,
cterm=self.CTerminus)
statsByRv, statGroupNames = self.stats_by_rv(data, RvSiteindexesMap,
genes, conditions,
interactions)
LogZPercByRep, NZMeanByRep = self.global_stats_for_rep(data)
self.transit_message("Running ZINB")
pvals, qvals, run_status = self.run_zinb(data, genes, NZMeanByRep,
LogZPercByRep,
RvSiteindexesMap, conditions,
covariates, interactions)
def orderStats(x, y):
ic1 = x.split(SEPARATOR)
ic2 = y.split(SEPARATOR)
c1, i1 = (ic1[0], ic1[1]) if len(ic1) > 1 else (ic1[0], None)
c2, i2 = (ic2[0], ic2[1]) if len(ic2) > 1 else (ic2[0], None)
if len(self.included_conditions) > 0:
condDiff = (self.included_conditions.index(c1) -
self.included_conditions.index(c2))
## Order by interaction, if stat belongs to same condition
if condDiff == 0 and i1 is not None and i2 is not None:
return (orderingMetadata['interaction'].index(i1) -
orderingMetadata['interaction'].index(i2))
return condDiff
## Order by samples metadata, if include flag not provided.
condDiff = (orderingMetadata['condition'].index(c1) -
orderingMetadata['condition'].index(c2))
if condDiff == 0 and i1 is not None and i2 is not None:
return (orderingMetadata['interaction'].index(i1) -
orderingMetadata['interaction'].index(i2))
return condDiff
orderedStatGroupNames = sorted(statGroupNames,
key=functools.cmp_to_key(orderStats))
headersStatGroupNames = [
x.replace(SEPARATOR, '_') for x in orderedStatGroupNames
]
self.transit_message("Adding File: %s" % (self.output))
file = open(self.output, "w")
if len(headersStatGroupNames) == 2: lfcNames = ["LFC"]
else: lfcNames = list(map(lambda v: "LFC_" + v, headersStatGroupNames))
head = ("Rv Gene TAs".split() +
list(map(lambda v: "Mean_" + v, headersStatGroupNames)) +
lfcNames +
list(map(lambda v: "NZmean_" + v, headersStatGroupNames)) +
list(map(lambda v: "NZperc_" + v, headersStatGroupNames)) +
"pval padj".split() + ["status"])
file.write("#Console: python3 %s\n" % " ".join(sys.argv))
file.write(
"#parameters: normalization=%s, trimming=%s/%s%% (N/C), pseudocounts=%s\n"
% (self.normalization, self.NTerminus, self.CTerminus, self.PC))
file.write('#' + '\t'.join(head) + EOL)
for gene in genes:
Rv = gene["rv"]
means = [
statsByRv[Rv]['mean'][group] for group in orderedStatGroupNames
]
PC = self.PC
if len(means) == 2:
LFCs = [numpy.math.log((means[1] + PC) / (means[0] + PC), 2)]
else:
m = numpy.mean(means)
LFCs = [numpy.math.log((x + PC) / (m + PC), 2) for x in means]
vals = ([Rv, gene["gene"],
str(len(RvSiteindexesMap[Rv]))] + [
"%0.1f" % statsByRv[Rv]['mean'][group]
for group in orderedStatGroupNames
] + ["%0.3f" % x for x in LFCs] + [
"%0.1f" % statsByRv[Rv]['nz_mean'][group]
for group in orderedStatGroupNames
] + [
"%0.2f" % statsByRv[Rv]['nz_perc'][group]
for group in orderedStatGroupNames
] + ["%f" % x
for x in [pvals[Rv], qvals[Rv]]]) + [run_status[Rv]]
file.write('\t'.join(vals) + EOL)
file.close()
self.transit_message("Finished Zinb analysis")
self.transit_message("Time: %0.1fs\n" % (time.time() - start_time))
def Run(self):
self.transit_message("Starting Gene Mean Counts Export")
start_time = time.time()
#Get orf data
self.transit_message("Getting Data")
if self.combined_wig:
(position, fulldata,
datasets) = tnseq_tools.read_combined_wig(self.ctrldata[0])
else:
(fulldata, position) = tnseq_tools.get_data(self.ctrldata)
(fulldata,
factors) = norm_tools.normalize_data(fulldata, self.normalization,
self.ctrldata,
self.annotation_path)
position = position.astype(int)
hash = transit_tools.get_pos_hash(self.annotation_path)
rv2info = transit_tools.get_gene_info(self.annotation_path)
self.transit_message("Normalizing")
self.output.write("#Summarized to Mean Gene Counts with TRANSIT.\n")
if self.normalization != "nonorm":
self.output.write("#Reads normalized using '%s'\n" %
self.normalization)
if type(factors[0]) == type(0.0):
self.output.write(
"#Normalization Factors: %s\n" %
"\t".join(["%s" % f for f in factors.flatten()]))
else:
self.output.write("#Normalization Factors: %s\n" % " ".join(
[",".join(["%s" % bx for bx in b]) for b in factors]))
self.output.write("#Files:\n")
names = datasets if self.combined_wig else self.ctrldata
for f in names:
self.output.write("#%s\n" % f)
K, Nsites = fulldata.shape
# Get Gene objects
if self.combined_wig:
G = tnseq_tools.Genes(self.ctrldata,
self.annotation_path,
norm=self.normalization,
data=fulldata,
position=position)
else:
G = tnseq_tools.Genes(self.ctrldata,
self.annotation_path,
norm=self.normalization)
N = len(G)
self.progress_range(N)
if self.combined_wig: dataset_header = '\t'.join(datasets)
else:
dataset_header = "\t".join(
[transit_tools.fetch_name(D) for D in self.ctrldata])
self.output.write("#Orf\tName\tNumber of TA sites\t%s\n" %
dataset_header)
for i, gene in enumerate(G):
if gene.n > 0:
data_str = "\t".join(
["%1.2f" % (M) for M in numpy.mean(gene.reads, 1)])
else:
data_str = "\t".join(["%1.2f" % (Z) for Z in numpy.zeros(K)])
self.output.write("%s\t%s\t%s\t%s\n" %
(gene.orf, gene.name, gene.n, data_str))
# Update progress
text = "Running Export Method... %5.1f%%" % (100.0 * i / N)
self.progress_update(text, i)
self.output.close()
self.transit_message("") # Printing empty line to flush stdout
self.finish()
self.transit_message("Finished Export")
def Run(self):
self.transit_message("Starting ZINB analysis")
start_time = time.time()
packnames = ("MASS", "pscl")
r_packages_needed = [x for x in packnames if not rpackages.isinstalled(x)]
if (len(r_packages_needed) > 0):
self.transit_error(
"Error: Following R packages are required: %(0)s. From R console, You can install them using install.packages(c(%(0)s))"
% ({'0': '"{0}"'.format('", "'.join(r_packages_needed))}))
sys.exit(1)
self.transit_message("Getting Data")
(sites, data, filenamesInCombWig) = tnseq_tools.read_combined_wig(self.combined_wig)
self.transit_message("Normalizing using: %s" % self.normalization)
(data, factors) = norm_tools.normalize_data(data, self.normalization)
condition_name = self.condition
conditionsByFile, covariatesByFileList, interactionsByFileList, orderingMetadata = tnseq_tools.read_samples_metadata(self.metadata, self.covars, self.interactions, condition_name=condition_name)
## [Condition] in the order of files in combined wig
conditions = self.wigs_to_conditions(
conditionsByFile,
filenamesInCombWig)
## [Covariate] in the order of files in combined wig
covariates = self.wigs_to_covariates(
covariatesByFileList,
filenamesInCombWig)
## [Interaction] in the order of files in combined wig
interactions = self.wigs_to_interactions(
interactionsByFileList,
filenamesInCombWig)
data, conditions, covariates, interactions = self.filter_wigs_by_conditions(
data,
conditions,
covariates = covariates,
interactions = interactions,
ignored_conditions = self.ignored_conditions,
included_conditions = self.included_conditions)
genes = tnseq_tools.read_genes(self.annotation_path)
TASiteindexMap = {TA: i for i, TA in enumerate(sites)}
RvSiteindexesMap = tnseq_tools.rv_siteindexes_map(genes, TASiteindexMap, nterm=self.NTerminus, cterm=self.CTerminus)
statsByRv, statGroupNames = self.stats_by_rv(data, RvSiteindexesMap, genes, conditions, interactions)
LogZPercByRep, NZMeanByRep = self.global_stats_for_rep(data)
self.transit_message("Running ZINB")
pvals, qvals, run_status = self.run_zinb(data, genes, NZMeanByRep, LogZPercByRep, RvSiteindexesMap, conditions, covariates, interactions)
def orderStats(x, y):
ic1 = x.split("_")
ic2 = y.split("_")
c1, i1 = (ic1[0], ic1[1]) if len(ic1) > 1 else (ic1[0], None)
c2, i2 = (ic2[0], ic2[1]) if len(ic2) > 1 else (ic2[0], None)
if len(self.included_conditions) > 0:
condDiff = (self.included_conditions.index(c1) - self.included_conditions.index(c2))
## Order by interaction, if stat belongs to same condition
if condDiff == 0 and i1 is not None and i2 is not None:
return (orderingMetadata['interaction'].index(i1) - orderingMetadata['interaction'].index(i2))
return condDiff
## Order by samples metadata, if include flag not provided.
condDiff = (orderingMetadata['condition'].index(c1) - orderingMetadata['condition'].index(c2))
if condDiff == 0 and i1 is not None and i2 is not None:
return (orderingMetadata['interaction'].index(i1) - orderingMetadata['interaction'].index(i2))
return condDiff
orderedStatGroupNames = sorted(statGroupNames, orderStats)
self.transit_message("Adding File: %s" % (self.output))
file = open(self.output,"w")
head = ("Rv Gene TAs".split() +
map(lambda v: "Mean_" + v, orderedStatGroupNames) +
map(lambda v: "NZmean_" + v, orderedStatGroupNames) +
map(lambda v: "NZperc_" + v, orderedStatGroupNames) +
"pval padj".split() + ["status"])
file.write("#Console: python %s\n" % " ".join(sys.argv))
file.write('\t'.join(head)+EOL)
for gene in genes:
Rv = gene["rv"]
vals = ([Rv, gene["gene"], str(len(RvSiteindexesMap[Rv]))] +
["%0.2f" % statsByRv[Rv]['mean'][group] for group in orderedStatGroupNames] +
["%0.2f" % statsByRv[Rv]['nz_mean'][group] for group in orderedStatGroupNames] +
["%0.2f" % statsByRv[Rv]['nz_perc'][group] for group in orderedStatGroupNames] +
["%f" % x for x in [pvals[Rv], qvals[Rv]]]) + [run_status[Rv]]
file.write('\t'.join(vals)+EOL)
file.close()
self.transit_message("Finished Zinb analysis")
self.transit_message("Time: %0.1fs\n" % (time.time() - start_time))
def Run(self):
self.transit_message("Starting ZINB analysis")
start_time = time.time()
packnames = ("MASS", "pscl")
r_packages_needed = [x for x in packnames if not rpackages.isinstalled(x)]
if (len(r_packages_needed) > 0):
self.transit_error(
"Error: Following R packages are required: %(0)s. From R console, You can install them using install.packages(c(%(0)s))"
% ({'0': '"{0}"'.format('", "'.join(r_packages_needed))}))
sys.exit(1)
self.transit_message("Getting Data")
(sites, data, filenamesInCombWig) = tnseq_tools.read_combined_wig(self.combined_wig)
self.transit_message("Normalizing using: %s" % self.normalization)
(data, factors) = norm_tools.normalize_data(data, self.normalization)
condition_name = self.condition
conditionsByFile, covariatesByFileList, interactionsByFileList, orderingMetadata = tnseq_tools.read_samples_metadata(self.metadata, self.covars, self.interactions, condition_name=condition_name)
## [Condition] in the order of files in combined wig
conditions = self.wigs_to_conditions(
conditionsByFile,
filenamesInCombWig)
## [Covariate] in the order of files in combined wig
covariates = self.wigs_to_covariates(
covariatesByFileList,
filenamesInCombWig)
## [Interaction] in the order of files in combined wig
interactions = self.wigs_to_interactions(
interactionsByFileList,
filenamesInCombWig)
data, conditions, covariates, interactions = self.filter_wigs_by_conditions(
data,
conditions,
covariates = covariates,
interactions = interactions,
ignored_conditions = self.ignored_conditions,
included_conditions = self.included_conditions)
genes = tnseq_tools.read_genes(self.annotation_path)
TASiteindexMap = {TA: i for i, TA in enumerate(sites)}
RvSiteindexesMap = tnseq_tools.rv_siteindexes_map(genes, TASiteindexMap, nterm=self.NTerminus, cterm=self.CTerminus)
statsByRv, statGroupNames = self.stats_by_rv(data, RvSiteindexesMap, genes, conditions, interactions)
LogZPercByRep, NZMeanByRep = self.global_stats_for_rep(data)
self.transit_message("Running ZINB")
pvals, qvals, run_status = self.run_zinb(data, genes, NZMeanByRep, LogZPercByRep, RvSiteindexesMap, conditions, covariates, interactions)
def orderStats(x, y):
ic1 = x.split(SEPARATOR)
ic2 = y.split(SEPARATOR)
c1, i1 = (ic1[0], ic1[1]) if len(ic1) > 1 else (ic1[0], None)
c2, i2 = (ic2[0], ic2[1]) if len(ic2) > 1 else (ic2[0], None)
if len(self.included_conditions) > 0:
condDiff = (self.included_conditions.index(c1) - self.included_conditions.index(c2))
## Order by interaction, if stat belongs to same condition
if condDiff == 0 and i1 is not None and i2 is not None:
return (orderingMetadata['interaction'].index(i1) - orderingMetadata['interaction'].index(i2))
return condDiff
## Order by samples metadata, if include flag not provided.
condDiff = (orderingMetadata['condition'].index(c1) - orderingMetadata['condition'].index(c2))
if condDiff == 0 and i1 is not None and i2 is not None:
return (orderingMetadata['interaction'].index(i1) - orderingMetadata['interaction'].index(i2))
return condDiff
orderedStatGroupNames = sorted(statGroupNames, key=functools.cmp_to_key(orderStats))
headersStatGroupNames = [x.replace(SEPARATOR,'_') for x in orderedStatGroupNames]
self.transit_message("Adding File: %s" % (self.output))
file = open(self.output,"w")
head = ("Rv Gene TAs".split() +
list(map(lambda v: "Mean_" + v, headersStatGroupNames)) +
list(map(lambda v: "NZmean_" + v, headersStatGroupNames)) +
list(map(lambda v: "NZperc_" + v, headersStatGroupNames)) +
"pval padj".split() + ["status"])
file.write("#Console: python %s\n" % " ".join(sys.argv))
file.write('\t'.join(head)+EOL)
for gene in genes:
Rv = gene["rv"]
vals = ([Rv, gene["gene"], str(len(RvSiteindexesMap[Rv]))] +
["%0.2f" % statsByRv[Rv]['mean'][group] for group in orderedStatGroupNames] +
["%0.2f" % statsByRv[Rv]['nz_mean'][group] for group in orderedStatGroupNames] +
["%0.2f" % statsByRv[Rv]['nz_perc'][group] for group in orderedStatGroupNames] +
["%f" % x for x in [pvals[Rv], qvals[Rv]]]) + [run_status[Rv]]
file.write('\t'.join(vals)+EOL)
file.close()
self.transit_message("Finished Zinb analysis")
self.transit_message("Time: %0.1fs\n" % (time.time() - start_time))