def correlationPairwise(outf, inf, group1, group2, method=pcorrelation):
"""Computes all pairwise correlations between group1 and group2"""
inData = mData.rCRSData(inf)
inRows = inData[inData.keys()[0]].keys()
f = open(outf, "w")
for i in group1:
if i not in inData:
continue
list1 = []
for k in inRows:
list1.append(inData[i][k])
for j in group2:
if j not in inData:
continue
list2 = []
for k in inRows:
list2.append(inData[j][k])
value = method(list1, list2)
f.write("%s\t%s\t%s\n" % (i, j, value))
f.close()
def correlationPairwise(outf, inf, group1, group2, method = pcorrelation):
"""Computes all pairwise correlations between group1 and group2"""
inData = mData.rCRSData(inf)
inRows = inData[inData.keys()[0]].keys()
f = open(outf, "w")
for i in group1:
if i not in inData:
continue
list1 = []
for k in inRows:
list1.append(inData[i][k])
for j in group2:
if j not in inData:
continue
list2 = []
for k in inRows:
list2.append(inData[j][k])
value = method(list1, list2)
f.write("%s\t%s\t%s\n" % (i, j, value))
f.close()
def correlationMatrix(outf, inf, method = pcorrelation):
"""Takes a tab file and cross-correlates the columns"""
outData = dict()
inData = mData.rCRSData(inf)
inCols = inData.keys()
inRows = inData[inCols[0]].keys()
for i in inCols:
outData[i] = dict()
outData[i][i] = 1.0
for i in range(len(inCols)-1):
list1 = []
for k in inRows:
list1.append(inData[inCols[i]][k])
for j in range(i+1, len(inCols)):
list2 = []
for k in inRows:
list2.append(inData[inCols[j]][k])
value = method(list1, list2)
outData[inCols[i]][inCols[j]] = value
outData[inCols[j]][inCols[i]] = value
mData.wCRSData(outf, outData)
def correlationMatrix(outf, inf, method=pcorrelation):
"""Takes a tab file and cross-correlates the columns"""
outData = dict()
inData = mData.rCRSData(inf)
inCols = inData.keys()
inRows = inData[inCols[0]].keys()
for i in inCols:
outData[i] = dict()
outData[i][i] = 1.0
for i in range(len(inCols) - 1):
list1 = []
for k in inRows:
list1.append(inData[inCols[i]][k])
for j in range(i + 1, len(inCols)):
list2 = []
for k in inRows:
list2.append(inData[inCols[j]][k])
value = method(list1, list2)
outData[inCols[i]][inCols[j]] = value
outData[inCols[j]][inCols[i]] = value
mData.wCRSData(outf, outData)
"%s\t%s\t%s\t%s\t%s\n"
% (
feature,
statsMap[feature][1],
",".join(membersMap[feature]),
proportionMap[feature],
statsMap[feature][0],
)
)
f.close()
## summarize pathways
if iplMatrix is not None:
pathwayVals = {}
pathwayScores = {}
(iplData, iplSamples, iplFeatures) = mData.rCRSData(iplMatrix, retFeatures=True)
f = open(re.split("/", sifFile)[-2] + ".members.output", "r")
for line in f:
line = line.rstrip("\n\r")
pline = re.split("\t", line)
if pline[1] in statsMap:
if pline[1] not in pathwayVals:
pathwayVals[pline[1]] = {}
for sample in iplSamples:
pathwayVals[pline[1]][sample] = []
if pline[0] in iplFeatures:
for sample in iplSamples:
pathwayVals[pline[1]][sample].append(abs(float(iplData[sample][pline[0]])))
f.close()
for pathwayName in pathwayVals.keys():
pathwayScores[pathwayName] = {}
elif o == "-q":
verbose = False
## execute
samples = []
features = []
if sampleFile != None:
samples = mData.rList(sampleFile)
if featureFile != None:
features = mData.rList(featureFile)
## read circleFiles
circleData = []
circleColors = []
for i in range(len(circleFiles)):
(data, cols, rows) = mData.rCRSData(circleFiles[i], retFeatures = True)
circleData.append(data)
minCol = rgb(0, 0, 255)
zerCol = rgb(255, 255, 255)
maxCol = rgb(255, 0, 0)
if circleFiles[i].endswith("meth"):
maxCol = rgb(0, 0, 255)
minCol = rgb(255, 0, 0)
log("Color: meth\n")
elif circleFiles[i].startswith("mut."):
maxCol = rgb(0, 0, 0)
minCol = rgb(255, 255, 255)
log("Color: mut\n")
circleColors.append( (minCol, zerCol, maxCol) )
if sampleFile == None:
samples = list(set(cols) | set(samples))
def cli_routine(outputDir, circleFiles, orderFiles, sampleFile, featureFile, orderFeature, centerFile, colorscaleFile, printLabel, verbose, cohortMinMax=False, purpleHack = True): """Routine for program execution via command-line.""" # I've tried not to touch this method as much as possible. # I don't want to break the way it was working for Sam Ng. # chrisw ## execute samples = [] features = [] if sampleFile != None: samples = mData.rList(sampleFile) if featureFile != None: features = mData.rList(featureFile) # end section for getting lists of samples and features ## read circleFiles # circleData is a list of dict[col][row]=score from each circleFile circleData = [] # circleColorsPalette is a list of (minColor),(zeroColor),(maxColor) circleColorsPalette = [] ## read colorscaleFile # the format is as follows - header compulsory: # min/max color coding color1 color2 color 3 # -2,2 rgb 155,155,155 255,255,255 0,0,0, # - rgb 155,0,155 255,0,255 0,0,0, # the "color format" is intended to support more color format, as I have # seen the html-colors in the code. # Michael ([email protected]) colorscaleData = None if colorscaleFile != None: if cohortMinMax: log("WARNING: The -k option overrides -m") colorscaleData = mData.retRows(colorscaleFile,aslist=True) line=1 for cs in colorscaleData: line = line + 1 if len(cs) != 5: log("ERROR: color scale needs five fields: datapoints, colorcoding(rgb) and three colors\n", die = True) try: cs[0] = [float(x) for x in cs[0].split(",")] except ValueError: pass if len(cs[0]) != 2 and cs[0] != "-": print cs[0] log("ERROR: Two data points or dash needed for color scale\n", die = True) if cs[1].lower() == "rgb": try: cs[2] = rgb(*[float(x) for x in cs[2].split(",")]) cs[3] = rgb(*[float(x) for x in cs[3].split(",")]) cs[4] = rgb(*[float(x) for x in cs[4].split(",")]) except TypeError: log("ERROR: RGB needs three values on line " + str(line) + "\n", die = True) except ValueError: log("ERROR: RGB color not correctly defined on line " + str(line) + "\n", die=True) else: log("ERROR: Unknown color coding on line " + str(line) + ": " + str(cs[1]) + "\n", die=True) for i in xrange(len(circleFiles)): # get data, samples, and features from each circleFile # data is a dict[col][row]=score # cols is a list of sample names # features is a list of feature names (data, cols, rows) = mData.rCRSData(circleFiles[i], retFeatures=True) circleData.append(data) minCol = lightBlueRGB zerCol = whiteRGB maxCol = redRGB if colorscaleFile != None and i<len(colorscaleData): #get colors from specified colorscaleFile minCol = colorscaleData[i][2] zerCol = colorscaleData[i][3] maxCol = colorscaleData[i][4] # special cases for -meth and -mut # if circleFiles[i].endswith("meth"): # maxCol = blueRGB # minCol = redRGB # log("Color: meth\n") # elif circleFiles[i].endswith("mut"): # maxCol = blackRGB # minCol = whiteRGB # log("Color: mut\n") circleColorsPalette.append((minCol, zerCol, maxCol)) # if no sampleFile/featureFile, default to using samples/features from circleFiles if sampleFile == None: samples = list(set(cols) | set(samples)) if featureFile == None: features = list(set(rows) | set(features)) # end section for reading circleFiles ## read centerFile centerData = None if centerFile != None: centerData = mData.r2Col(centerFile, header=True) ## sort if orderFeature != None: if len(orderFiles) > 0: orderData = [] orderColors = [] for i in xrange(len(orderFiles)): orderData.append(mData.rCRSData(orderFiles[i])) minCol = whiteRGB zerCol = whiteRGB maxCol = blackRGB orderColors.append((minCol, zerCol, maxCol)) else: orderData = circleData # sort samples based on sample score in orderData # priority of sorting determined by orderFiles parameter samples.sort(lambda x, y: scmp(x, y, orderFeature, orderData)) ## cohort png # cgi will probably not use orderFiles if len(orderFiles) > 0: imgFile = "%s/Cohort.png" % (outputDir) label = "Cohort" centerCol = whiteRGB.tohex() cohortCircleCols = [] for i in xrange(len(orderData)): ringCols = [] ringVals = [] for sample in samples: if sample in orderData[i]: if orderFeature in orderData[i][sample]: ringVals.append(orderData[i][sample][orderFeature]) elif "*" in orderData[i][sample]: ringVals.append(orderData[i][sample]["*"]) minVal = min([-0.01] + mData.floatList(ringVals)) maxVal = max([0.01] + mData.floatList(ringVals)) for sample in samples: if sample in orderData[i]: if orderFeature in orderData[i][sample]: ringCols.append(getColor(orderData[i][sample][orderFeature], minVal, maxVal, minColor=orderColors[i][0], zeroColor=orderColors[i][1], maxColor=orderColors[i][2])) elif "*" in orderData[i][sample]: ringCols.append(getColor(orderData[i][sample]["*"], minVal, maxVal, minColor=orderColors[i][0], zeroColor=orderColors[i][1], maxColor=orderColors[i][2])) else: ringCols.append(greyRGB.tohex()) else: ringCols.append(greyRGB.tohex()) cohortCircleCols.append(ringCols) plotCircle(imgFile, label=label, centerCol=centerCol, circleCols=cohortCircleCols, innerRadTotal=0.2, outerRadTotal=0.5, width=5) # end section for sample ordering ## plot images if centerData != None: centerDataFloatList = mData.floatList(centerData.values()) centerDataMinVal = min([-0.01] + centerDataFloatList) centerDataMaxVal = max([0.01] + centerDataFloatList) # get min/max values for datasets if cohortMinMax: (minValList, maxValList) = getCohortMinMaxValues(features, samples, circleData) else: (minValList, maxValList) = (None, None) if colorscaleData != None: (minValList, maxValList) = getColorScaleMinMaxValues(minValList, maxValList, len(circleData), colorscaleData) for feature in features: log("Drawing %s\n" % (feature)) centerColHex = None if centerData != None: if feature in centerData: centerColHex = getColor(centerData[feature], centerDataMinVal, centerDataMaxVal, minColor=lightBlueRGB, zeroColor=whiteRGB, purple0Hack=purpleHack) imgFile = "%s/%s.png" % (outputDir, re.sub("[/:]", "_", feature)) label = "" if printLabel: label = feature image_width = 5.0 drawCircleImageForFeature(feature, samples, label, imgFile, circleData, circleColorsPalette, width=image_width, centerColHex=centerColHex, minValList=minValList, maxValList=maxValList, purple0Hack=purpleHack) for sample in samples: log("ordered samples: %s\n" % (sample))
if i.endswith(".html"):
continue
elif i.endswith(".py"):
continue
elif i.endswith("js"):
continue
elif i.endswith("swf"):
continue
else:
log("Working %s ...\n" % (i))
d.write(htmlIndexItem % ("%s.html" % (i), i))
f = open("%s.html" % (i), "w")
f.write(htmlHead)
f.write(htmlCategory % i)
if os.path.exists("%s/stats.tab" % (i)):
(samData, samCols, samItems) = mData.rCRSData("%s/stats.tab" % (i), retFeatures = True)
samCols = ["Link"] + samCols
else:
samData = {}
samCols = ["Link", "Note"]
for j in samCols:
samData[j] = {}
samItems = []
for j in os.listdir(i):
if j.startswith("img"):
continue
samItems.append(re.sub(".html", "", j))
for j in samItems:
samData["Link"][j] = "%s.html" % (j)
samData["Note"][j] = ""
f.write(htmlTableHead % ("</th>\n <th>".join(samCols)))
def filterNet(files, phenotypes = [], statLine = None, outDir = None):
global filterBounds
filterString = "%s_%s" % (filterBounds[0], filterBounds[1])
## read global pathway
(gNodes, gInteractions) = mPathway.rPathway(globalPathway)
## read drugs
#drugData = mData.rSet(drugBank)
## write LABEL.NA, TYPE.NA
if outputAttributes:
typef = open("TYPE.NA", "w")
labelf = open("LABEL.NA", "w")
typef.write("TYPE (class=java.lang.String)\n")
labelf.write("LABEL (class=java.lang.String)\n")
for i in gNodes.keys():
typef.write("%s = %s\n" % (i, gNodes[i]))
if gNodes[i] == "protein":
labelf.write("%s = %s\n" % (i, i))
else:
labelf.write("%s = %s\n" % (i, ""))
#drugs here
typef.close()
labelf.close()
## read scores
uData = dict()
sData = dict()
for i in range(len(files)):
uData[i] = mData.rCRSData(files[i])
sData[i] = dict()
for j in uData[i].keys():
sData[i][j] = dict()
for k in uData[i][j].keys():
try:
sData[i][j][k] = abs(float(uData[i][j][k]))
except ValueError:
sData[i][j][k] = "NA"
## iterate phenotypes
for p in sData[0].keys():
if len(phenotypes) > 0:
if p not in phenotypes:
continue
pNodes = dict()
pInteractions = dict()
## write SCORE.NA
if outputAttributes:
scoref = open(p+"_SCORE.NA", "w")
scoref.write("SCORE (class=java.lang.Float)\n")
for i in gNodes.keys():
if i in uData[0][p]:
if uData[0][p][i] == "NA":
scoref.write("%s = %s\n" % (i, "0"))
else:
scoref.write("%s = %s\n" % (i, uData[0][p][i]))
else:
scoref.write("%s = %s\n" % (i, "0"))
scoref.close()
## compute thresholds
pStats = []
if statLine == None:
for i in range(len(sData.keys())):
pStats.append(mCalculate.mean_std(sData[i][p].values()))
else:
for i in re.split(",",statLine):
(v1, v2) = re.split(";",i)
pStats.append((float(v1), float(v2)))
log("%s\t%s;%s" % (p, pStats[0][0], pStats[0][1]))
for i in range(1, len(pStats)):
log(",%s;%s" % (pStats[i][0], pStats[i][1]))
log("\n")
## iterate links
for a in gInteractions.keys():
if a not in sData[0][p]:
continue
elif sData[0][p][a] == "NA":
continue
for b in gInteractions[a].keys():
if b not in sData[0][p]:
continue
elif sData[0][p][b] == "NA":
continue
## score nodes by threshold
aScore = []
bScore = []
linkScore = []
for i in range(len(sData.keys())):
linkScore.append([sData[i][p][a], sData[i][p][b]])
for i in range(len(sData.keys())):
if linkScore[i][0] > pStats[i][0]+filterBounds[1]*pStats[i][1]:
aScore.append(2)
elif linkScore[i][0] > pStats[i][0]+filterBounds[0]*pStats[i][1]:
aScore.append(1)
else:
aScore.append(0)
if linkScore[i][1] > pStats[i][0]+filterBounds[1]*pStats[i][1]:
bScore.append(2)
elif linkScore[i][1] > pStats[i][0]+filterBounds[0]*pStats[i][1]:
bScore.append(1)
else:
bScore.append(0)
## selection rule
if includeType == "OR":
if max(aScore)+max(bScore) >= 3:
(pNodes, pInteractions) = addLink(a, b, pNodes, pInteractions, gNodes, gInteractions)
elif includeType == "AND":
votes = 0
for i in range(len(sData.keys())):
if aScore[i]+bScore[i] >= 3:
votes += 0
if votes == len(sData.keys()):
(pNodes, pInteractions) = addLink(a, b, pNodes, pInteractions, gNodes, gInteractions)
elif includeType == "MAIN":
if aScore[0]+bScore[0] >= 3:
(pNodes, pInteractions) = addLink(a, b, pNodes, pInteractions, gNodes, gInteractions)
## connect top scoring disconnected nodes
sortedTop = []
for i in sData[0][p].keys():
if i not in gNodes:
continue
if gNodes[i] in ["protein"]:
sortedTop.append(i)
sortedTop.sort(lambda x, y: cmp(sData[0][p][y],sData[0][p][x]))
while (sData[0][p][sortedTop[0]] == "NA"):
sortedTop.pop(0)
if len(sortedTop) == 0:
break
for i in range(topDisconnected):
if i > len(sortedTop)-1:
break
if sData[0][p][sortedTop[i]] < pStats[0][0]+filterBounds[0]*pStats[0][1]:
break
if sortedTop[i] not in gNodes:
continue
if sortedTop[i] not in pNodes:
pNodes[sortedTop[i]] = gNodes[sortedTop[i]]
pInteractions[sortedTop[i]] = dict()
pInteractions[sortedTop[i]]["__DISCONNECTED__"] = "-disconnected-"
## output
if outDir == None:
wrtDir = p
else:
wrtDir = outDir
if not os.path.exists(wrtDir):
os.system("mkdir %s" % (wrtDir))
## output for pathway-predictor
if outputPARADIGM:
protSet = set()
for i in gNodes:
if gNodes[i] == "protein":
protSet.update([i])
netNodes = mPathway.sortConnected(pNodes, pInteractions, mPathway.revInteractions(pInteractions))
trainNodes = []
for i in netNodes:
if len((protSet) & set(i)) > featureReq:
trainNodes += i
if len(trainNodes) == 0:
log("ERROR: no nets contained enough data\n...trying again\n")
if filterBounds[0]+0.1 <= filterBounds[1]:
filterBounds[1] -= 0.1
else:
filterBounds[0] -= 0.1
filterBounds[1] -= 0.1
filterNet(files, phenotypes = phenotypes, statLine = statLine, outDir = outDir)
sys.exit(0)
(lNodes, lInteractions) = mPathway.constructInteractions(trainNodes, pNodes, pInteractions)
if outputAttributes:
mPathway.wSIF("%s/%s_%s_pp.sif" % (wrtDir, p, filterString), lInteractions)
## connect class node
classNode = "class"
lInteractions[classNode] = dict()
for i in lNodes.keys():
if i not in protSet:
continue
lInteractions[classNode][i] = "-cl>"
lNodes[classNode] = "active"
mPathway.wPathway("%s/%s_%s_pp.tab" % (wrtDir, p, filterString), lNodes, lInteractions)
## output nodrug pathway
else:
mPathway.wSIF("%s/%s_%s_nodrug.sif" % (wrtDir, p, filterString), pInteractions)
(cpNodes, cpInteractions) = mPathway.filterComplexesByGeneSupport(pNodes, pInteractions,
mPathway.revInteractions(pInteractions), gNodes,
mPathway.getComponentMap(gNodes, mPathway.revInteractions(gInteractions)))
mPathway.wSIF("%s/%s_%s_nodrug_cleaned.sif" % (wrtDir, p, filterString), cpInteractions)
## output summary
f = open(re.split("/", sifFile)[-2] + ".summary.tab", "w")
features = statsMap.keys()
features.sort(lambda x, y: cmp(statsMap[x][1], statsMap[y][1]))
for feature in features:
f.write("%s\t%s\t%s\t%s\t%s\n" %
(feature, statsMap[feature][1], ",".join(membersMap[feature]),
proportionMap[feature], statsMap[feature][0]))
f.close()
## summarize pathways
if iplMatrix is not None:
pathwayVals = {}
pathwayScores = {}
(iplData, iplSamples, iplFeatures) = mData.rCRSData(iplMatrix,
retFeatures=True)
f = open(re.split("/", sifFile)[-2] + ".members.output", "r")
for line in f:
line = line.rstrip("\n\r")
pline = re.split("\t", line)
if pline[1] in statsMap:
if pline[1] not in pathwayVals:
pathwayVals[pline[1]] = {}
for sample in iplSamples:
pathwayVals[pline[1]][sample] = []
if pline[0] in iplFeatures:
for sample in iplSamples:
pathwayVals[pline[1]][sample].append(
abs(float(iplData[sample][pline[0]])))
f.close()
for pathwayName in pathwayVals.keys():
continue
elif i.endswith(".py"):
continue
elif i.endswith("js"):
continue
elif i.endswith("swf"):
continue
else:
log("Working %s ...\n" % (i))
d.write(htmlIndexItem % ("%s.html" % (i), i))
f = open("%s.html" % (i), "w")
f.write(htmlHead)
f.write(htmlCategory % i)
if os.path.exists("%s/stats.tab" % (i)):
(samData, samCols,
samItems) = mData.rCRSData("%s/stats.tab" % (i),
retFeatures=True)
samCols = ["Link"] + samCols
else:
samData = {}
samCols = ["Link", "Note"]
for j in samCols:
samData[j] = {}
samItems = []
for j in os.listdir(i):
if j.startswith("img"):
continue
samItems.append(re.sub(".html", "", j))
for j in samItems:
samData["Link"][j] = "%s.html" % (j)
samData["Note"][j] = ""
f.write(htmlTableHead % ("</th>\n <th>".join(samCols)))
def cli_routine(outputDir,
circleFiles,
orderFiles,
sampleFile,
featureFile,
orderFeature,
centerFile,
colorscaleFile,
printLabel,
verbose,
cohortMinMax=False,
purpleHack=True):
"""Routine for program execution via command-line."""
# I've tried not to touch this method as much as possible.
# I don't want to break the way it was working for Sam Ng.
# chrisw
## execute
samples = []
features = []
if sampleFile != None:
samples = mData.rList(sampleFile)
if featureFile != None:
features = mData.rList(featureFile)
# end section for getting lists of samples and features
## read circleFiles
# circleData is a list of dict[col][row]=score from each circleFile
circleData = []
# circleColorsPalette is a list of (minColor),(zeroColor),(maxColor)
circleColorsPalette = []
## read colorscaleFile
# the format is as follows - header compulsory:
# min/max color coding color1 color2 color 3
# -2,2 rgb 155,155,155 255,255,255 0,0,0,
# - rgb 155,0,155 255,0,255 0,0,0,
# the "color format" is intended to support more color format, as I have
# seen the html-colors in the code.
# Michael ([email protected])
colorscaleData = None
if colorscaleFile != None:
if cohortMinMax:
log("WARNING: The -k option overrides -m")
colorscaleData = mData.retRows(colorscaleFile, aslist=True)
line = 1
for cs in colorscaleData:
line = line + 1
if len(cs) != 5:
log("ERROR: color scale needs five fields: datapoints, colorcoding(rgb) and three colors\n",
die=True)
try:
cs[0] = [float(x) for x in cs[0].split(",")]
except ValueError:
pass
if len(cs[0]) != 2 and cs[0] != "-":
print cs[0]
log("ERROR: Two data points or dash needed for color scale\n",
die=True)
if cs[1].lower() == "rgb":
try:
cs[2] = rgb(*[float(x) for x in cs[2].split(",")])
cs[3] = rgb(*[float(x) for x in cs[3].split(",")])
cs[4] = rgb(*[float(x) for x in cs[4].split(",")])
except TypeError:
log("ERROR: RGB needs three values on line " + str(line) +
"\n",
die=True)
except ValueError:
log("ERROR: RGB color not correctly defined on line " +
str(line) + "\n",
die=True)
else:
log("ERROR: Unknown color coding on line " + str(line) + ": " +
str(cs[1]) + "\n",
die=True)
for i in xrange(len(circleFiles)):
# get data, samples, and features from each circleFile
# data is a dict[col][row]=score
# cols is a list of sample names
# features is a list of feature names
(data, cols, rows) = mData.rCRSData(circleFiles[i], retFeatures=True)
circleData.append(data)
minCol = lightBlueRGB
zerCol = whiteRGB
maxCol = redRGB
if colorscaleFile != None and i < len(colorscaleData):
#get colors from specified colorscaleFile
minCol = colorscaleData[i][2]
zerCol = colorscaleData[i][3]
maxCol = colorscaleData[i][4]
# special cases for -meth and -mut
# if circleFiles[i].endswith("meth"):
# maxCol = blueRGB
# minCol = redRGB
# log("Color: meth\n")
# elif circleFiles[i].endswith("mut"):
# maxCol = blackRGB
# minCol = whiteRGB
# log("Color: mut\n")
circleColorsPalette.append((minCol, zerCol, maxCol))
# if no sampleFile/featureFile, default to using samples/features from circleFiles
if sampleFile == None:
samples = list(set(cols) | set(samples))
if featureFile == None:
features = list(set(rows) | set(features))
# end section for reading circleFiles
## read centerFile
centerData = None
if centerFile != None:
centerData = mData.r2Col(centerFile, header=True)
## sort
if orderFeature != None:
if len(orderFiles) > 0:
orderData = []
orderColors = []
for i in xrange(len(orderFiles)):
orderData.append(mData.rCRSData(orderFiles[i]))
minCol = whiteRGB
zerCol = whiteRGB
maxCol = blackRGB
orderColors.append((minCol, zerCol, maxCol))
else:
orderData = circleData
# sort samples based on sample score in orderData
# priority of sorting determined by orderFiles parameter
samples.sort(lambda x, y: scmp(x, y, orderFeature, orderData))
## cohort png
# cgi will probably not use orderFiles
if len(orderFiles) > 0:
imgFile = "%s/Cohort.png" % (outputDir)
label = "Cohort"
centerCol = whiteRGB.tohex()
cohortCircleCols = []
for i in xrange(len(orderData)):
ringCols = []
ringVals = []
for sample in samples:
if sample in orderData[i]:
if orderFeature in orderData[i][sample]:
ringVals.append(orderData[i][sample][orderFeature])
elif "*" in orderData[i][sample]:
ringVals.append(orderData[i][sample]["*"])
minVal = min([-0.01] + mData.floatList(ringVals))
maxVal = max([0.01] + mData.floatList(ringVals))
for sample in samples:
if sample in orderData[i]:
if orderFeature in orderData[i][sample]:
ringCols.append(
getColor(orderData[i][sample][orderFeature],
minVal,
maxVal,
minColor=orderColors[i][0],
zeroColor=orderColors[i][1],
maxColor=orderColors[i][2]))
elif "*" in orderData[i][sample]:
ringCols.append(
getColor(orderData[i][sample]["*"],
minVal,
maxVal,
minColor=orderColors[i][0],
zeroColor=orderColors[i][1],
maxColor=orderColors[i][2]))
else:
ringCols.append(greyRGB.tohex())
else:
ringCols.append(greyRGB.tohex())
cohortCircleCols.append(ringCols)
plotCircle(imgFile,
label=label,
centerCol=centerCol,
circleCols=cohortCircleCols,
innerRadTotal=0.2,
outerRadTotal=0.5,
width=5)
# end section for sample ordering
## plot images
if centerData != None:
centerDataFloatList = mData.floatList(centerData.values())
centerDataMinVal = min([-0.01] + centerDataFloatList)
centerDataMaxVal = max([0.01] + centerDataFloatList)
# get min/max values for datasets
if cohortMinMax:
(minValList,
maxValList) = getCohortMinMaxValues(features, samples, circleData)
else:
(minValList, maxValList) = (None, None)
if colorscaleData != None:
(minValList,
maxValList) = getColorScaleMinMaxValues(minValList, maxValList,
len(circleData),
colorscaleData)
for feature in features:
log("Drawing %s\n" % (feature))
centerColHex = None
if centerData != None:
if feature in centerData:
centerColHex = getColor(centerData[feature],
centerDataMinVal,
centerDataMaxVal,
minColor=lightBlueRGB,
zeroColor=whiteRGB,
purple0Hack=purpleHack)
imgFile = "%s/%s.png" % (outputDir, re.sub("[/:]", "_", feature))
label = ""
if printLabel:
label = feature
image_width = 5.0
drawCircleImageForFeature(feature,
samples,
label,
imgFile,
circleData,
circleColorsPalette,
width=image_width,
centerColHex=centerColHex,
minValList=minValList,
maxValList=maxValList,
purple0Hack=purpleHack)
for sample in samples:
log("ordered samples: %s\n" % (sample))
log("ERROR: incorrect number of arguments", die=True)
phenotypeFile = args[0]
dataFile = args[1]
global verbose
for o, a in opts:
if o == "-q":
verbose = False
## execute
phenotypeName = re.split("/", phenotypeFile)[-1].rstrip(".tab")
dataName = re.split("/", dataFile)[-1].rstrip(".tab")
outputDir = "OCCAM__%s__%s" % (phenotypeName, dataName)
syscmd("mkdir %s" % (outputDir))
(phenData, phenColumns, phenRows) = mData.rCRSData(phenotypeFile,
retFeatures=True)
(matData, matColumns, matRows) = mData.rCRSData(dataFile, retFeatures=True)
## samples
posSamples = []
negSamples = []
for sample in phenRows:
if sample not in matColumns:
continue
if phenData[phenColumns[phenIndex]][sample] == "+":
posSamples.append(sample)
elif phenData[phenColumns[phenIndex]][sample] == "-":
negSamples.append(sample)
## output
f = open("%s/results.tab" % (outputDir), "w")
log("ERROR: incorrect number of arguments", die = True)
phenotypeFile = args[0]
dataFile = args[1]
global verbose
for o, a in opts:
if o == "-q":
verbose = False
## execute
phenotypeName = re.split("/", phenotypeFile)[-1].rstrip(".tab")
dataName = re.split("/", dataFile)[-1].rstrip(".tab")
outputDir = "OCCAM__%s__%s" % (phenotypeName, dataName)
syscmd("mkdir %s" % (outputDir))
(phenData, phenColumns, phenRows) = mData.rCRSData(phenotypeFile, retFeatures = True)
(matData, matColumns, matRows) = mData.rCRSData(dataFile, retFeatures = True)
## samples
posSamples = []
negSamples = []
for sample in phenRows:
if sample not in matColumns:
continue
if phenData[phenColumns[phenIndex]][sample] == "+":
posSamples.append(sample)
elif phenData[phenColumns[phenIndex]][sample] == "-":
negSamples.append(sample)
## output
f = open("%s/results.tab" % (outputDir), "w")
usage(1)
if (len(args) == 1):
tabFile = sys.stdin
htmlFile = args[0]
else:
tabFile = args[0]
htmlFile = args[1]
global verbose
for o, a in opts:
if o == "-q":
verbose = False
## read tabFile
(tabData, tabCols, tabRows) = mData.rCRSData(tabFile, retFeatures = True)
## write htmlFile
f = open("%s" % (htmlFile), "w")
f.write(htmlHead)
tabCols = ["id"] + tabCols
f.write(htmlTableHead % ("</th>\n <th>".join(tabCols)))
for i in tabRows:
tabList = []
for j in tabCols[1:]:
tabList.append(str(tabData[j][i]))
f.write(htmlTableItem % (i, "</td>\n <td>".join(tabList)))
f.write(htmlTableTail)
f.write(htmlTail)
f.close()
if sampleFile != None:
samples = mData.rList(sampleFile)
if featureFile != None:
features = mData.rList(featureFile)
## read circleFiles
circleData = []
circleColors = []
##
## record file types for each, effects the color scheme
## use the input index for each
color_scheme_map = {}
for i in range(len(circleFiles)):
circleFile, colorScheme = circleFiles[i].split(':')
color_scheme_map[i] = parseColorScheme(colorScheme)
(data, cols, rows) = mData.rCRSData(circleFile, retFeatures = True)
circleData.append(data)
#circleColors.append( (minCol, zerCol, maxCol) )
#if sampleFile == None:
# samples = list(set(cols) | set(samples))
#if featureFile == None:
# features = list(set(rows) | set(features))
mutationData = None
if mutationsFile != None:
mutationData, cols, rows = mData.rCRSData(mutationsFile, retFeatures = True)
## read centerFile
centerData = None
if centerFile != None:
centerData = mData.r2Col(centerFile, header = True)
