def __init__(self,
parent,
dataset_list=["H37Rv_Sassetti_glycerol.wig"],
annotation="H37Rv.prot_table",
gene="",
scale=None,
feature_hashes=[],
feature_data=[]):
view_trash.MainFrame.__init__(self, parent)
self.parent = parent
self.size = wx.Size(1500, 800)
self.start = 1
self.end = 10000
#self.orf2data = draw_trash.read_prot_table(annotation)
#self.hash = draw_trash.hash_prot_genes(annotation)
self.orf2data = transit_tools.get_gene_info(annotation)
self.hash = transit_tools.get_pos_hash(annotation)
self.features = []
#Data to facilitate search
self.name2id = {}
for orf, (name, desc, start, end, strand) in self.orf2data.items():
name = name.lower()
if name not in self.name2id: self.name2id[name] = []
self.name2id[name].append(orf)
self.lowerid2id = dict([(x.lower(), x) for x in self.orf2data.keys()])
self.labels = [fetch_name(d) for d in dataset_list]
(self.fulldata, self.position) = tnseq_tools.get_data(dataset_list)
#Save normalized data
(self.fulldata_norm,
self.factors) = norm_tools.normalize_data(self.fulldata,
method="nzmean")
self.wasNorm = False
#initialize parent class
self.feature_hashes = feature_hashes
self.feature_data = feature_data
if not scale:
scale = [150] * len(dataset_list)
self.scale = scale
self.globalScale = False
self.datasetChoice.SetItems(self.labels)
self.datasetChoice.SetSelection(0)
if gene:
self.searchText.SetValue(gene)
self.searchFunc(gene)
self.updateFunc(parent)
self.Fit()
def Run(self):
self.transit_message("Starting IGV Export")
start_time = time.time()
#Get orf data
self.transit_message("Getting Data")
(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("#Converted to IGV 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")
for f in self.ctrldata:
self.output.write("#%s\n" % f)
dataset_str = "\t".join(
[transit_tools.fetch_name(F) for F in self.ctrldata])
self.output.write("#Chromosome\tStart\tEnd\tFeature\t%s\tTAs\n" %
dataset_str)
chrom = transit_tools.fetch_name(self.annotation_path)
(K, N) = fulldata.shape
self.progress_range(N)
for i, pos in enumerate(position):
self.output.write(
"%s\t%s\t%s\tTA%s\t%s\t1\n" %
(chrom, position[i], position[i] + 1, position[i], "\t".join(
["%1.1f" % fulldata[j][i] for j in range(len(fulldata))])))
# 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 Gene Mean Counts Export")
start_time = time.time()
#Get orf data
self.transit_message("Getting Data")
(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")
for f in self.ctrldata:
self.output.write("#%s\n" % f)
K,Nsites = fulldata.shape
# Get Gene objects
G = tnseq_tools.Genes(self.ctrldata, self.annotation_path, norm=self.normalization)
N = len(G)
self.progress_range(N)
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 Combined Wig Export")
start_time = time.time()
#Get orf data
self.transit_message("Getting Data")
(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("#Converted to CombinedWig with TRANSIT.\n")
self.output.write("#normalization method: %s\n" % self.normalization)
if self.normalization != "nonorm":
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]))
(K, N) = fulldata.shape
for f in self.ctrldata:
self.output.write("#File: %s\n" % f)
self.output.write("#TAcoord\t%s\n" % ('\t'.join(self.ctrldata)))
for i, pos in enumerate(position):
#self.output.write("%d\t%s\t%s\n" % (position[i], "\t".join(["%1.1f" % c for c in fulldata[:,i]]),",".join(["%s (%s)" % (orf,rv2info.get(orf,["-"])[0]) for orf in hash.get(position[i], [])]) ))
if self.normalization != 'nonorm':
vals = "\t".join(["%1.1f" % c for c in fulldata[:, i]])
else:
vals = "\t".join(["%d" % c for c in fulldata[:, i]
]) # no decimals if raw counts
self.output.write("%d\t%s\t%s\n" % (position[i], vals, ",".join([
"%s (%s)" % (orf, rv2info.get(orf, ["-"])[0])
for orf in hash.get(position[i], [])
])))
# 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 IGV Export")
start_time = time.time()
#Get orf data
self.transit_message("Getting Data")
(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("#Converted to IGV 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")
for f in self.ctrldata:
self.output.write("#%s\n" % f)
dataset_str = "\t".join([transit_tools.fetch_name(F) for F in self.ctrldata])
self.output.write("#Chromosome\tStart\tEnd\tFeature\t%s\tTAs\n" % dataset_str)
chrom = transit_tools.fetch_name(self.annotation_path)
(K,N) = fulldata.shape
self.progress_range(N)
for i,pos in enumerate(position):
self.output.write("%s\t%s\t%s\tTA%s\t%s\t1\n" % (chrom, position[i], position[i]+1, position[i], "\t".join(["%1.1f" % fulldata[j][i] for j in range(len(fulldata))])))
# 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 Combined Wig Export")
start_time = time.time()
#Get orf data
self.transit_message("Getting Data")
(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("#Converted to CombinedWig with TRANSIT.\n")
self.output.write("#normalization method: %s\n" % self.normalization)
if self.normalization != "nonorm":
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]))
(K,N) = fulldata.shape
for f in self.ctrldata:
self.output.write("#File: %s\n" % f)
self.output.write("#TAcoord\t%s\n" % ('\t'.join(self.ctrldata)))
for i,pos in enumerate(position):
#self.output.write("%d\t%s\t%s\n" % (position[i], "\t".join(["%1.1f" % c for c in fulldata[:,i]]),",".join(["%s (%s)" % (orf,rv2info.get(orf,["-"])[0]) for orf in hash.get(position[i], [])]) ))
if self.normalization!='nonorm': vals = "\t".join(["%1.1f" % c for c in fulldata[:,i]])
else: vals = "\t".join(["%d" % c for c in fulldata[:,i]]) # no decimals if raw counts
self.output.write("%d\t%s\t%s\n" % (position[i],vals,",".join(["%s (%s)" % (orf,rv2info.get(orf,["-"])[0]) for orf in hash.get(position[i], [])]) ))
# 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 addFeatureFunc(self, event):
wc = u"Known Annotation Formats (*.prot_table,*.gff3,*.gff)|*.prot_table;*.gff3;*.gff;|\nProt Table (*.prot_table)|*.prot_table;|\nGFF3 (*.gff,*.gff3)|*.gff;*.gff3;|\nAll files (*.*)|*.*"
path = self.OpenFile(DIR=self.parent.workdir, FILE="", WC=wc)
try:
if path:
if self.checkHMMFeature(path):
H, S = self.getHMMHash(path)
self.feature_hashes.append(H)
self.feature_data.append(S)
else:
self.feature_hashes.append(
transit_tools.get_pos_hash(path))
self.feature_data.append(transit_tools.get_gene_info(path))
self.updateFunc(self.parent)
self.Fit()
else:
self.track_message("No feature added")
except Exception as e:
self.track_message("ERROR: %s" % e)
traceback.print_exc()
def Run(self):
self.transit_message("Starting HMM Method")
start_time = time.time()
#Get data
self.transit_message("Getting Data")
(data, position) = transit_tools.get_validated_data(self.ctrldata,
wxobj=self.wxobj)
(K, N) = data.shape
# Normalize data
if 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)
# Do LOESS
if self.LOESS:
self.transit_message("Performing LOESS Correction")
for j in range(K):
data[j] = stat_tools.loess_correction(position, data[j])
hash = transit_tools.get_pos_hash(self.annotation_path)
rv2info = transit_tools.get_gene_info(self.annotation_path)
if len(self.ctrldata) > 1:
self.transit_message("Combining Replicates as '%s'" %
self.replicates)
O = tnseq_tools.combine_replicates(
data, method=self.replicates
) + 1 # Adding 1 to because of shifted geometric in scipy
#Parameters
Nstates = 4
label = {0: "ES", 1: "GD", 2: "NE", 3: "GA"}
reads = O - 1
reads_nz = sorted(reads[reads != 0])
size = len(reads_nz)
mean_r = numpy.average(reads_nz[:int(0.95 * size)])
mu = numpy.array([1 / 0.99, 0.01 * mean_r + 2, mean_r, mean_r * 5.0])
#mu = numpy.array([1/0.99, 0.1 * mean_r + 2, mean_r, mean_r*5.0])
L = 1.0 / mu
B = [] # Emission Probability Distributions
for i in range(Nstates):
B.append(scipy.stats.geom(L[i]).pmf)
pins = self.calculate_pins(O - 1)
pins_obs = sum([1 for rd in O if rd >= 2]) / float(len(O))
pnon = 1.0 - pins
pnon_obs = 1.0 - pins_obs
for r in range(100):
if pnon**r < 0.01: break
A = numpy.zeros((Nstates, Nstates))
a = math.log1p(-B[int(Nstates / 2)](1)**r)
b = r * math.log(B[int(Nstates / 2)](1)) + math.log(
1.0 / 3) # change to Nstates-1?
for i in range(Nstates):
A[i] = [b] * Nstates
A[i][i] = a
PI = numpy.zeros(Nstates) # Initial state distribution
PI[0] = 0.7
PI[1:] = 0.3 / (Nstates - 1)
self.progress_range(self.maxiterations)
###############
### VITERBI ###
(Q_opt, delta, Q) = self.viterbi(A, B, PI, O)
###############
##################
### ALPHA PASS ###
(log_Prob_Obs, alpha,
C) = self.forward_procedure(numpy.exp(A), B, PI, O)
##################
#################
### BETA PASS ###
beta = self.backward_procedure(numpy.exp(A), B, PI, O, C)
#################
T = len(O)
total = 0
state2count = dict.fromkeys(range(Nstates), 0)
for t in range(T):
state = Q_opt[t]
state2count[state] += 1
total += 1
self.output.write("#HMM - Sites\n")
self.output.write("# Tn-HMM\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("# \n")
self.output.write("# Mean:\t%2.2f\n" % (numpy.average(reads_nz)))
self.output.write("# Median:\t%2.2f\n" % numpy.median(reads_nz))
self.output.write("# Normalization:\t%s\n" % self.normalization)
self.output.write("# LOESS Correction:\t%s\n" % str(self.LOESS))
self.output.write("# pins (obs):\t%f\n" % pins_obs)
self.output.write("# pins (est):\t%f\n" % pins)
self.output.write("# Run length (r):\t%d\n" % r)
self.output.write("# State means:\n")
self.output.write("# %s\n" % " ".join(
["%s: %8.4f" % (label[i], mu[i]) for i in range(Nstates)]))
self.output.write("# Self-Transition Prob:\n")
self.output.write("# %s\n" % " ".join(
["%s: %2.4e" % (label[i], A[i][i]) for i in range(Nstates)]))
self.output.write("# State Emission Parameters (theta):\n")
self.output.write("# %s\n" % " ".join(
["%s: %1.4f" % (label[i], L[i]) for i in range(Nstates)]))
self.output.write("# State Distributions:")
self.output.write("# %s\n" % " ".join([
"%s: %2.2f%%" % (label[i], state2count[i] * 100.0 / total)
for i in range(Nstates)
]))
states = [int(Q_opt[t]) for t in range(T)]
last_orf = ""
for t in range(T):
s_lab = label.get(states[t], "Unknown State")
gamma_t = (alpha[:, t] * beta[:, t]) / numpy.sum(
alpha[:, t] * beta[:, t])
genes_at_site = hash.get(position[t], [""])
genestr = ""
if not (len(genes_at_site) == 1 and not genes_at_site[0]):
genestr = ",".join([
"%s_(%s)" % (g, rv2info.get(g, "-")[0])
for g in genes_at_site
])
self.output.write("%s\t%s\t%s\t%s\t%s\n" %
(int(position[t]), int(O[t]) - 1, "\t".join(
["%-9.2e" % g
for g in gamma_t]), s_lab, genestr))
self.output.close()
self.transit_message("") # Printing empty line to flush stdout
self.transit_message("Finished HMM - Sites Method")
self.transit_message("Adding File: %s" % (self.output.name))
self.add_file(filetype="HMM - Sites")
#Gene Files
self.transit_message("Creating HMM Genes Level Output")
genes_path = ".".join(self.output.name.split(
".")[:-1]) + "_genes." + self.output.name.split(".")[-1]
tempObs = numpy.zeros((1, len(O)))
tempObs[0, :] = O - 1
self.post_process_genes(tempObs, position, states, genes_path)
self.transit_message("Adding File: %s" % (genes_path))
self.add_file(path=genes_path, filetype="HMM - Genes")
self.finish()
self.transit_message("Finished HMM Method")