def makeResults(self, parallelResult, oEnvironment, args):
allDiffRegData = {}
allDiffRegSims = defaultdict(dict)
conditions = set()
for file in args.diffreg:
thisData = EnrichmentDF(DataFrame.parseFromFile(file))
condPair = tuple(thisData.getConditions())
for cond in condPair:
conditions.add(cond)
allDiffRegData[condPair] = thisData
for method in args.methods:
methodFCs = []
for x in thisData.getColumn(method + "_log2FC"):
if x != None and x!= 'None':
methodFCs.append(abs(float(x)))
average = sum(methodFCs) / len(methodFCs)
allDiffRegSims[method][condPair] = average
allConditions = sorted(list(conditions))
for method in allDiffRegSims:
sims = np.zeros( (len(allConditions), len(allConditions)) )
for condPair in allDiffRegSims[method]:
sims[ allConditions.index(condPair[0]), allConditions.index(condPair[1]) ] = allDiffRegSims[method][condPair]
sims[allConditions.index(condPair[1]), allConditions.index(condPair[0])] = allDiffRegSims[method][
condPair]
PorePlot.heat_map_cluster(sims, allConditions, allConditions, "Similarity: " + str(method), "", pltcfg=args.pltcfg)
def __init__(self, args):
super(FoldChangeAnalysis, self).__init__(args)
self.counts = None
self.condData = EnrichmentDF()
def makeResults(self, parallelResult, oEnvironment, args):
if not args.counts == None:
"""
counts is a defaultdict(list) for each condition name with maybe multiple samples
"""
counts = self.readCounts(args)
vConds = sorted([x for x in counts])
createdComparisons = defaultdict(list)
conditions = []
for valueSource in ['coverage', 'read_counts']:
self.condData = EnrichmentDF()
replicates = {}
for condition in vConds:
condData = counts[condition]
condReplicates = []
for condDataSample in condData:
geneNames = condDataSample.getColumnIndex('gene')
geneCounts = condDataSample.getColumnIndex(valueSource)
condRow = condDataSample.toDataRow(
geneNames, geneCounts)
sampleName = condDataSample.filepath
conditions.append(sampleName)
condReplicates.append(sampleName)
self.condData.addCondition(condRow, sampleName)
replicates[condition] = condReplicates
print("Running for conditions: " + str(vConds))
createdComparisons[valueSource] += self.condData.runDEanalysis(
args.output,
prefix=valueSource,
rscriptPath=args.rscript.name,
methods=args.methods,
replicates=replicates,
noDErun=args.noanalysis)
self.prepareHTMLOut(createdComparisons, replicates, args)
if args.diffreg != None:
createdComparisons = defaultdict(list)
conditions = set()
for file in args.diffreg:
df = EnrichmentDF.parseFromFile(file)
valueSource = self.getValueSource(df)
conditions += df.getConditions()
createdComparisons[valueSource].append(file)
self.prepareHTMLOut(createdComparisons, conditions, args)
def makeResults(self, parallelResult, oEnvironment, args):
allDiffRegData = {}
allDiffRegSims = defaultdict(dict)
conditions = set()
def parseNones(row):
ret = [None] * len(row)
for i in range(0, len(row)):
if row[i] != 'None':
ret[i] = row[i]
return ret
topGenes = Counter()
for file in args.diffreg:
thisData = EnrichmentDF(DataFrame.parseFromFile(file))
thisData.applyToRow(parseNones)
condPair = tuple(thisData.getConditions())
for cond in condPair:
conditions.add(cond)
allDiffRegData[condPair] = thisData
for method in args.methods:
methodFCs = []
pvals = thisData.toDataRow(
thisData.getColumnIndex('id'),
thisData.getColumnIndex(method + "_RAW.PVA"))
genepval = [(x[0], float(x[1])) for x in pvals.to_pairs()
if x[1] != None]
genepval.sort(key=lambda x: x[1])
for i in range(0, args.top):
topGenes[genepval[i][0]] += 1
outputDF = DataFrame()
geneIDidx = outputDF.addColumn('gene_id')
countIdx = outputDF.addColumn('count')
linkIdx = outputDF.addColumn('link')
for (gene, count) in topGenes.most_common():
geneRow = DataRow.fromDict({
'gene_id':
gene,
'count':
count,
'link':
"<a href='http://www.uniprot.org/uniprot/?query=" + gene +
"&sort=score' target='_blank'>UniProt</a>",
})
outputDF.addRow(geneRow)
outputDF.export(args.output, ExportTYPE.HTML)
def makeResults(self, parallelResult, oEnvironment, args):
if not args.counts == None:
"""
counts is a defaultdict(list) for each condition name with maybe multiple samples
"""
geneEnhancement = self.loadEnhancement(args.enhanced)
geneLengths = self.loadGeneLengths(args.lengths)
counts, cond2samples = self.readCounts(args,
biotypes=geneEnhancement,
gene2length=geneLengths)
#vConds = sorted([x for x in counts])
vConds = [x for x in counts]
createdComparisons = defaultdict(list)
conditions = []
for valueSource in ['count']:
self.condData = EnrichmentDF()
replicates = OrderedDict()
for condition in vConds:
condData = counts[condition]
condReplicates = []
for condDataSample in condData:
geneNames = condDataSample.getColumnIndex('gene')
geneCounts = condDataSample.getColumnIndex(valueSource)
rowUpdates = []
sampleName = condDataSample.filepath
print(sampleName, len(condDataSample))
for row in condDataSample:
rowData = {
"id": row["gene"],
sampleName: row[valueSource]
}
if args.libsize:
rowData[sampleName + ".LS"] = row["LS"]
if args.fpkm:
rowData[sampleName + ".FPKM"] = row["FPKM"]
if args.tpm:
rowData[sampleName + ".TPM"] = row["TPM"]
rowUpdates.append(rowData)
#condRows = condDataSample.namedRows(geneNames, interestCols)
#condRow = condDataSample.toDataRow(geneNames, geneCounts)
conditions.append(sampleName)
condReplicates.append(sampleName)
print("Add Condition", sampleName, rowUpdates[0])
self.condData.addConditions(rowUpdates, sampleName)
replicates[condition] = condReplicates
print("Running for conditions: " + str(vConds))
createdComparisons[valueSource] += self.condData.runDEanalysis(
args.output,
prefix=valueSource,
rscriptPath=args.rscript.name,
methods=args.methods,
replicates=replicates,
noDErun=args.noanalysis,
enhanceSymbol=geneEnhancement,
geneLengths=geneLengths)
self.prepareHTMLOut(createdComparisons, replicates, args)
def __init__(self, args):
super(FoldChangeSimilarity, self).__init__(args)
self.counts = None
self.condData = EnrichmentDF()