def objListUtility_mafBlockCounts(data, mb, featLen, numBins):
""" Utility function for the MafBlock instance version of
libMafGffPlot.objListToBinnedWiggle()
This is by far the most costly routine in the objList creation process
"""
from libMafGffPlot import objListUtility_rangeToPos
import numpy
length = mb.refEnd - (mb.refStart + 1)
# tough to follow index hack to get around the fact that numpy will not use
# += 1 for a list of indices that contain repeats.
pos = objListUtility_rangeToPos(mb.refStart, mb.refEnd, featLen, numBins)
plo, phi = pos.min(), pos.max()
pbins = numpy.bincount(pos - plo)
data['maf'][plo:phi + 1] += pbins
for i in xrange(2, 8):
if length >= 10**i:
data['maf1e%d' % i][plo:phi + 1] += pbins
if mb.spl >= 10**i:
data['mafSpl1e%d' % i][plo:phi + 1] += pbins
if mb.pairTotalLength >= 10**i:
data['mafCtg1e%d' % i][plo:phi + 1] += pbins
if mb.hpl >= 10**i:
data['mafCpl1e%d' % i][plo:phi + 1] += pbins
def objListUtility_mafBlockCounts(data, mb, featLen, numBins):
""" Utility function for the MafBlock instance version of
libMafGffPlot.objListToBinnedWiggle()
This is by far the most costly routine in the objList creation process
"""
from libMafGffPlot import objListUtility_rangeToPos
import numpy
length = mb.refEnd - (mb.refStart + 1)
# tough to follow index hack to get around the fact that numpy will not use
# += 1 for a list of indices that contain repeats.
pos = objListUtility_rangeToPos(mb.refStart, mb.refEnd, featLen, numBins)
plo, phi = pos.min(), pos.max()
pbins = numpy.bincount(pos - plo)
data["maf"][plo : phi + 1] += pbins
for i in xrange(2, 8):
if length >= 10 ** i:
data["maf1e%d" % i][plo : phi + 1] += pbins
if mb.spl >= 10 ** i:
data["mafSpl1e%d" % i][plo : phi + 1] += pbins
if mb.pairTotalLength >= 10 ** i:
data["mafCtg1e%d" % i][plo : phi + 1] += pbins
if mb.hpl >= 10 ** i:
data["mafCpl1e%d" % i][plo : phi + 1] += pbins
def objListToBinnedWiggle(objList, featLen, numBins, filename):
""" obj can be either a GffRecord object or a MafBlock object.
featLen is the length of the chromosome.
returns a numpy vector of length numBins normalized by the maximum
possible number of bases per bin.
"""
from libMafGffPlot import GffRecord
from libMafGffPlot import MafBlock
from libMafGffPlot import newMafWigDict
from libMafGffPlot import objListUtility_xAxis
import numpy
import sys
if objList is None or len(objList) < 1:
return None
if isinstance(objList[0], GffRecord):
""" the Gff return is a single numpy vector of numBins length
"""
data = {}
# populate xAxis
data['xAxis'] = objListUtility_xAxis(featLen, numBins)
annotTypes = set(
['CDS', 'UTR', 'NXE', 'NGE', 'island', 'tandem', 'repeat'])
for t in annotTypes:
data[t + 'Count'] = numpy.zeros(shape=(numBins))
data[t + 'Max'] = 0
for a in objList:
if a.type not in annotTypes:
continue
# verify input
if a.start > featLen or a.end > featLen:
sys.stderr.write(
'libMafGffPlot.py: file %s has annotation on chr %s '
'with bounds [%d - %d] which are beyond featLen (%d)\n' %
(filename, a.chr, a.start, a.end, featLen))
sys.exit(1)
# index position in a 'numBins' length array.
pos = objListUtility_rangeToPos(a.start, a.end, featLen, numBins)
# tough to follow index hack to get around the fact that numpy will not use
# += 1 for a list of indices that contain repeats.
plo, phi = pos.min(), pos.max()
pbins = numpy.bincount(pos - plo)
data[a.type + 'Count'][plo:phi + 1] += pbins
for p in pos:
if data[a.type + 'Max'] < data[a.type + 'Count'][p]:
data[a.type + 'Max'] = data[a.type + 'Count'][p]
return data
elif isinstance(objList[0], MafBlock):
""" the Maf return is a dictionary with the following keys
maf all maf block bases
maf1e2 maf blocks 100 or greater
maf1e3 maf blocks 1,000 or greater
maf1e4 maf blocks 10,000 or greater
maf1e5 maf blocks 100,000 or greater
maf1e6 maf blocks 1,000,000 or greater
maf1e7 maf blocks 10,000,000 or greater
xAxis x Values
mafCpl1eX maf contig paths of X or greater
mafCtg1eX maf contigs of X or greater. taken from totalLength field of maf.
mafSpl1eX maf scaffold paths of X or greater
mafCpEdgeCounts each contig path has two edges, a left and a right
mafCpEdgeMax max count
mafCpErrorCounts contig paths are made up of segments, segments may have errors at junctions.
mafCpErrorMax max count
mafSpEdgeCounts Same as above, but for scaffold paths
mafSpEdgeMax
mafSpErrorCounts
mafSpErrorMax
blockEdgeCounts each block has two edges, a left and a right
blockEdgeMax max count
"""
from libMafGffPlot import objListUtility_addContigPathEdgeErrors
from libMafGffPlot import objListUtility_addBlockEdges
from libMafGffPlot import objListUtility_normalizeCategories
data = newMafWigDict(numBins)
# populate xAxis
data['xAxis'] = objListUtility_xAxis(featLen, numBins)
for mb in objList:
# do block edges
objListUtility_addBlockEdges(data, mb, featLen, numBins)
# do contige path edges and errors
objListUtility_addContigPathEdgeErrors(data, mb, featLen, numBins)
# do all of the different maf block flavors
objListUtility_mafBlockCounts(data, mb, featLen, numBins)
# normalize all categories
objListUtility_normalizeCategories(data, featLen, numBins)
return data
# closing the elif isinstance() checks
else:
return None
def objListToBinnedWiggle(objList, featLen, numBins, filename):
""" obj can be either a GffRecord object or a MafBlock object.
featLen is the length of the chromosome.
returns a numpy vector of length numBins normalized by the maximum
possible number of bases per bin.
"""
from libMafGffPlot import GffRecord
from libMafGffPlot import MafBlock
from libMafGffPlot import newMafWigDict
from libMafGffPlot import objListUtility_xAxis
import numpy
import sys
if objList is None or len(objList) < 1:
return None
if isinstance(objList[0], GffRecord):
""" the Gff return is a single numpy vector of numBins length
"""
data = {}
# populate xAxis
data["xAxis"] = objListUtility_xAxis(featLen, numBins)
annotTypes = set(["CDS", "UTR", "NXE", "NGE", "island", "tandem", "repeat"])
for t in annotTypes:
data[t + "Count"] = numpy.zeros(shape=(numBins))
data[t + "Max"] = 0
for a in objList:
if a.type not in annotTypes:
continue
# verify input
if a.start > featLen or a.end > featLen:
sys.stderr.write(
"libMafGffPlot.py: file %s has annotation on chr %s "
"with bounds [%d - %d] which are beyond featLen (%d)\n" % (filename, a.chr, a.start, a.end, featLen)
)
sys.exit(1)
# index position in a 'numBins' length array.
pos = objListUtility_rangeToPos(a.start, a.end, featLen, numBins)
# tough to follow index hack to get around the fact that numpy will not use
# += 1 for a list of indices that contain repeats.
plo, phi = pos.min(), pos.max()
pbins = numpy.bincount(pos - plo)
data[a.type + "Count"][plo : phi + 1] += pbins
for p in pos:
if data[a.type + "Max"] < data[a.type + "Count"][p]:
data[a.type + "Max"] = data[a.type + "Count"][p]
return data
elif isinstance(objList[0], MafBlock):
""" the Maf return is a dictionary with the following keys
maf all maf block bases
maf1e2 maf blocks 100 or greater
maf1e3 maf blocks 1,000 or greater
maf1e4 maf blocks 10,000 or greater
maf1e5 maf blocks 100,000 or greater
maf1e6 maf blocks 1,000,000 or greater
maf1e7 maf blocks 10,000,000 or greater
xAxis x Values
mafCpl1eX maf contig paths of X or greater
mafCtg1eX maf contigs of X or greater. taken from totalLength field of maf.
mafSpl1eX maf scaffold paths of X or greater
mafCpEdgeCounts each contig path has two edges, a left and a right
mafCpEdgeMax max count
mafCpErrorCounts contig paths are made up of segments, segments may have errors at junctions.
mafCpErrorMax max count
mafSpEdgeCounts Same as above, but for scaffold paths
mafSpEdgeMax
mafSpErrorCounts
mafSpErrorMax
blockEdgeCounts each block has two edges, a left and a right
blockEdgeMax max count
"""
from libMafGffPlot import objListUtility_addContigPathEdgeErrors
from libMafGffPlot import objListUtility_addBlockEdges
from libMafGffPlot import objListUtility_normalizeCategories
data = newMafWigDict(numBins)
# populate xAxis
data["xAxis"] = objListUtility_xAxis(featLen, numBins)
for mb in objList:
# do block edges
objListUtility_addBlockEdges(data, mb, featLen, numBins)
# do contige path edges and errors
objListUtility_addContigPathEdgeErrors(data, mb, featLen, numBins)
# do all of the different maf block flavors
objListUtility_mafBlockCounts(data, mb, featLen, numBins)
# normalize all categories
objListUtility_normalizeCategories(data, featLen, numBins)
return data
# closing the elif isinstance() checks
else:
return None
