def mapRewriteInTbOld(genome_fID, tandemGapMax=0):
nbTandemDup = 0
# the rewritten genome
# Need to keep dictionnaries because there is often a gap in chromosome notations
tmp_genome_tb = {}
tb2g = {}
# Number of Genic Tandem Duplication global
for c in genome_fID:
tb2g[c] = []
# temp values
tmp_genome_tb[c] = []
old_fID = 'foo' # Not None, since None is used to mark genes to be removed
for i, (fID, s) in enumerate(genome_fID[c]):
# if the current gene is not a paralog of the previous gene
if fID != old_fID:
# start a new TB
tmp_genome_tb[c].append(fID)
# add the current gene index to the new tb component of tb2g
tb2g[c].append([i])
else:
# fID == old_fID:
# The index of the current gene is added to the last tb
tb2g[c][-1].append(i)
nbTandemDup += 1
old_fID = fID
# TODO be sure that this step is consistent with the 2D distance metric chosen.
# For instance, if the DPD is chosen, on vertical and horizontal lines, the
# distances are not consistent with the 1D distance metric.
if tandemGapMax > 0:
nbTandemDup = 0
tandemDistMax = tandemGapMax + 1
#TODO next loops could be optimised
tmp_tb2g = {}
combinator = myTools.myCombinator()
for c, chrom_tb in tmp_genome_tb.iteritems():
# print >> sys.stderr, "Length in tbs before tandem gap = %s" % len(chrom_tb)
for (i, fID) in enumerate(chrom_tb):
# Range starts at 2 because since genomes are already rewritten
# in tbs there is no adjacent tbs.
isAlone = True
for dist in range(2, min(tandemDistMax + 1, len(chrom_tb) - i)):
if chrom_tb[i + dist].n == chrom_tb[i].n:
pair = (i + dist, i)
# Add a link between the elements of a pair
combinator.addLink(pair)
isAlone = False
if isAlone:
combinator.addLink((i, i))
combinator.reduce()
tbChains = list(combinator)
# print >> sys.stderr, "Nb of chains of at least 1 tb = %s" % len(tbChains)
# print >> sys.stderr, "Nb of chains of at least 2 tbs = %s" % len([a for a in tbChains if len(a) >=2])
# print >> sys.stderr, "Chains of at least 2 tbs = %s" % [a for a in tbChains if len(a) >= 2]
# Sort neighbourhood by the increasing smallest index of the neighbourhood.
# TODO List could be yielded sorted (improve the myCombinator class)
for tbChain in tbChains:
tbChain.sort()
#print >> sys.stderr, "len(tbChains)=%s" % len(tbChains)
# Since what precedes the next script line is equivalent to:
# chainsOfTbs.sort(lambda x: min(x))
tbChains.sort(key=lambda x: x[0])
firstChainTbIdxs = []
otherChainTbIdxs = []
for tbChain in tbChains:
if len(tbChain) == 1:
firstChainTbIdx = tbChain[0]
otherChainTbIdxs.append([])
elif len(tbChain) >= 2:
firstChainTbIdx = tbChain[0]
otherChainTbIdxs.append([])
for tbIdx in tbChain[1:]:
otherChainTbIdxs[-1].append(tbIdx)
else:
raise
firstChainTbIdxs.append(firstChainTbIdx)
assert len(firstChainTbIdxs) == len(otherChainTbIdxs)
tmp_tb2g[c] = []
for (firstTb, otherTbs) in zip(firstChainTbIdxs, otherChainTbIdxs):
tmp_tb2g[c].append([])
for oldTbIdx in [firstTb] + otherTbs:
# lis1 + list2 returns the concatenation of the two lists
tmp_tb2g[c][-1] = tmp_tb2g[c][-1] + tb2g[c][oldTbIdx]
assert len(tmp_tb2g[c]) == len(firstChainTbIdxs), len(tmp_tb2g[c])
nbTandemDup +=\
sum([len(gTandemDuplicates)-1 for gTandemDuplicates in tmp_tb2g[c]])
# DEBUG assertion
listIsSorted = lambda l: all([i] <= l[i+1] for i in range(len(l)-1))
assert listIsSorted(tmp_tb2g[c])
combinator.reset()
tb2g = tmp_tb2g
mtb2g = {}
for (c, newMapC) in tb2g.iteritems():
mtb2g[c] = Mapping(newMapC)
# DEBUG assertion
# nbOffIDGenes = sum([len(chrom) for chrom in genome_fID.values()])
# nbOfTbs = sum([len(chrom) for chrom in genome_tb.values()])
# assert N_GTD_g == nbOffIDGenes - nbOfTbs, "%s == %s - %s" % (N_GTD_g, nbOffIDGenes, nbOfTbs)
return (mtb2g, (nbTandemDup))
def mapRewriteInTb(genome_fID, tandemGapMax=0):
"""
:param genome_fID: [..., (fID,s), ...] with 'fID' the line number of the gene in the ancGene ans 's' the strand of the gene in the genome
:param tandemGapMax:
:return: mtb2g : a mapping corresponding to the rewritting process
"""
assert isinstance(genome_fID, myLightGenomes.LightGenome)
# the rewritten genome
# Need to keep dictionaries because there is often a gap in chromosome notations
tb2g = {}
# TODO be sure that this step is consistent with the 2D distance metric chosen.
# For instance, if the DPD is chosen, on vertical and horizontal lines, the
# distances are not consistent with the 1D distance metric.
nbTandemDup = 0
tandemDistMax = tandemGapMax + 1
#TODO next loops could be optimised
combinator = myTools.myCombinator()
for c, chrom_tb in genome_fID.iteritems():
tb2g[c] = []
# print >> sys.stderr, "Length in tbs before tandem gap = %s" % len(chrom_tb)
for (i, fID) in enumerate(chrom_tb):
isAlone = True
for dist in range(1, min(tandemDistMax + 1, len(chrom_tb) - i)):
if chrom_tb[i + dist].n == chrom_tb[i].n:
pair = (i + dist, i)
# Add a link between the elements of a pair
combinator.addLink(pair)
isAlone = False
if isAlone:
combinator.addLink((i, i))
combinator.reduce()
tbChains = list(combinator)
# print >> sys.stderr, "Nb of chains of at least 1 tb = %s" % len(tbChains)
# print >> sys.stderr, "Nb of chains of at least 2 tbs = %s" % len([a for a in tbChains if len(a) >=2])
# print >> sys.stderr, "Chains of at least 2 tbs = %s" % [a for a in tbChains if len(a) >= 2]
# Sort neighbourhood by the increasing smallest index of the neighbourhood.
# TODO List could be yielded sorted (improve the myCombinator class)
for tbChain in tbChains:
tbChain.sort()
#print >> sys.stderr, "len(tbChains)=%s" % len(tbChains)
# Since what precedes the next script line is equivalent to:
# chainsOfTbs.sort(lambda x: min(x))
tbChains.sort(key=lambda x: x[0])
firstChainTbIdxs = []
otherChainTbIdxs = []
for tbChain in tbChains:
if len(tbChain) == 1:
firstChainTbIdx = tbChain[0]
otherChainTbIdxs.append([])
elif len(tbChain) >= 2:
firstChainTbIdx = tbChain[0]
otherChainTbIdxs.append([])
for tbIdx in tbChain[1:]:
otherChainTbIdxs[-1].append(tbIdx)
else:
raise
firstChainTbIdxs.append(firstChainTbIdx)
assert len(firstChainTbIdxs) == len(otherChainTbIdxs)
for (firstTb, otherTbs) in zip(firstChainTbIdxs, otherChainTbIdxs):
tb2g[c].append([])
for oldTbIdx in [firstTb] + otherTbs:
# lis1 + list2 returns the concatenation of the two lists
tb2g[c][-1].append(oldTbIdx)
assert len(tb2g[c]) == len(firstChainTbIdxs), len(tb2g[c])
nbTandemDup +=\
sum([len(gTandemDuplicates)-1 for gTandemDuplicates in tb2g[c]])
# DEBUG assertion
listIsSorted = lambda l: all([i] <= l[i+1] for i in range(len(l)-1))
assert listIsSorted(tb2g[c])
combinator.reset()
mtb2g = {}
for (c, newMapC) in tb2g.iteritems():
mtb2g[c] = Mapping(newMapC)
# #DEBUG assertion
# nbOffIDGenes = sum([len(chrom) for chrom in genome_fID.values()])
# nbOfTbs = sum([len(chrom) for chrom in tb2g.values()])
# assert nbTandemDup == nbOffIDGenes - nbOfTbs, "%s == %s - %s" % (nbTandemDup, nbOffIDGenes, nbOfTbs)
return (mtb2g, (nbTandemDup))