def stepwise_pileup_generator(index_positions: Iterator[IndexPosition],
alignment_file: pysam.AlignmentFile,
) -> Iterator[MotifPileup]:
"""MotifPileup Generator"""
# pylint: disable=stop-iteration-return
# (pylint false positive)
idx_pos_iterable = iter(index_positions)
curr_idx = next(idx_pos_iterable)
pileup_columns = alignment_file.pileup(reference=curr_idx.chrom,
start=curr_idx.start,
end=None,
truncate=True)
# Functional form required
# https://github.com/python/mypy/issues/4349
# noinspection PyPep8Naming
EmptyPileupColumn = NamedTuple('EmptyPileupColumn', [('reference_pos', int)])
empty_pileup_column = EmptyPileupColumn(-1)
empty_pileup_column_iterable = repeat(empty_pileup_column)
pileup_columns_iterator = chain(pileup_columns,
empty_pileup_column_iterable)
curr_pileup_column = next(pileup_columns_iterator)
curr_pileup_pos = curr_pileup_column.reference_pos
while True:
try:
if curr_pileup_pos == -1:
yield MotifPileup(reads=[], idx_pos=curr_idx)
for curr_idx in idx_pos_iterable:
yield MotifPileup(reads=[], idx_pos=curr_idx)
break
elif curr_idx.start > curr_pileup_pos:
curr_pileup_column = next(pileup_columns_iterator)
curr_pileup_pos = curr_pileup_column.reference_pos
continue
elif curr_idx.start < curr_pileup_pos:
yield MotifPileup(reads=[], idx_pos=curr_idx)
curr_idx = next(idx_pos_iterable)
continue
elif curr_idx.start == curr_pileup_pos:
pileup_reads = pileups(curr_pileup_column, curr_idx.watson_base)
yield MotifPileup(reads=pileup_reads, idx_pos=curr_idx)
curr_idx = next(idx_pos_iterable)
curr_pileup_column = next(pileup_columns_iterator)
curr_pileup_pos = curr_pileup_column.reference_pos
except StopIteration:
return # generator will now raise StopIteration
def data_generator_pysam(my_args, name, start, stop, is_bulk):
fasta_file = FastaFile(my_args.fasta)
ref = fasta_file.fetch(name, start, stop)
my_arg = {
'fastafile': fasta_file,
'stepper': 'samtools',
'adjust_capq_threshold': 50,
'contig': name,
'start': start,
'stop': stop,
'min_mapping_quality': 0 if is_bulk else 20,
'min_base_quality': 13,
}
if is_bulk:
bam_file = AlignmentFile(my_args.bulk, 'rb')
else:
bam_file = AlignmentFile(my_args.bam, 'rb')
read_bases_list = []
for pileup_column in bam_file.pileup(**my_arg):
pos = pileup_column.reference_pos
if pos >= stop:
break
if pos < start:
continue
read_bases_list = pileup_column.get_query_sequences(mark_matches=True,
mark_ends=True,
add_indels=True)
read_bases = ''.join(read_bases_list).upper()
n = pileup_column.get_num_aligned()
if n == 0:
read_bases = '*'
base_q = '*'
map_q = '*'
else:
base_q = ''.join([chr(int(i) + PHREDSCORE) \
for i in pileup_column.get_query_qualities()])
map_q = ''.join([chr(int(i) + PHREDSCORE) \
for i in pileup_column.get_mapping_qualities()])
yield [name, pos, ref[pos - start], str(n), read_bases, base_q, map_q]
yield None
def analyzeAlignment(self, alignmentOutputDirectory):
print ('\nStep 2.) Parse the alignment and create a new consensus sequence.')
# Load up the Alignment Reference file, we'll need it.
alignmentReferenceFileName = join(alignmentOutputDirectory,'AlignmentReference.fasta')
alignmentRef = list(parse(alignmentReferenceFileName, 'fasta'))[0]
# Count the reads in the input file
totalReadCount = len(list(parse(self.readInput, self.readInputFormat)))
#self.readInputFormat
#self.readInput
# We generate a new consensus sequence from the alignment results.
newConsensusSequence = ""
# Open the bam file
bamfile = AlignmentFile(join(alignmentOutputDirectory,'alignment.bam'), 'rb')
# Open alignment analysis text file
alignmentSummaryFile = createOutputFile(join(alignmentOutputDirectory,'AlignmentSummary.csv'))
alignmentSummaryFile.write('Ref_Position,Ref_Base,Reference_Adjustment,Aligned_Count,Unaligned_Count,Match_Count,Mismatch_Count,In_Count,Del_Count,A_Count,G_Count,C_Count,T_Count\n')
# A smaller log. I will provide human-readable descriptions of the
# bases that were adjusted in the new consensus sequence.
# TODO: Provide surrounding sequence as well, maybe it's a repeat region....
# Acutally NAH, I want to just put it in the wrangler log.
#adjustedBasesSummaryFile = createOutputFile(join(alignmentOutputDirectory,'AdjustedBases.txt'))
# Todo: I should keep a more structured array of info for these alignments.
# Store this info into an object
#class columnStats():
alignmentInfo = AlignmentInfo()
# Keep a running total of adjustments made to the reference.
# If this total is 0, then theoretically the consensus matches the alignment reference, and we're done.
totalSequenceAdjustments = 0
# Iterate the reference sequence column by column.
pileupIterator = bamfile.pileup(alignmentRef.id)
for pileupColumn in pileupIterator:
currentAlignmentColumn = AlignmentColumn()
#columnResults = None
# columnResults.name='ll'
#
"""referencePosition = 0
referenceBase = ''
referenceAdjustment = '?'
alignedCount = 0
unalignedCount = 0
matchCount = 0
mismatchCount = 0
inCount = 0
delCount = 0
aCount = 0
gCount = 0
cCount = 0
tCount = 0"""
currentAlignmentColumn.referencePosition = pileupColumn.reference_pos
currentAlignmentColumn.referenceBase = alignmentRef[pileupColumn.reference_pos].upper()
currentAlignmentColumn.alignedCount = pileupColumn.nsegments
currentAlignmentColumn.unalignedCount = totalReadCount - currentAlignmentColumn.alignedCount
# Iterate the Reads at this position
for pileupRead in pileupColumn.pileups:
# If this read is a deletion
if(pileupRead.is_del == 1):
currentAlignmentColumn.delCount += 1
# else if this read is an insertion
elif(pileupRead.indel > 0):
#print ('INSERTION DETECTED, INDEL=' + str(pileupRead.indel))
currentAlignmentColumn.inCount += 1
# Else if it is a refskip (TODO What does this mean? no read aligned? Count these?)
elif(pileupRead.is_refskip):
print('This read is a refskip, i dont know what that means:' + pileupRead.alignment.query_name)
raise Exception('This read is a refskip, i dont know what that means:' + pileupRead.alignment.query_name)
# else this means we have a base aligned at this position for this read.
else:
currentBase = pileupRead.alignment.query_sequence[pileupRead.query_position].upper()
#print('Reference,Current:' + referenceBase + ',' + currentBase)
#print('Curr')
if(currentBase == currentAlignmentColumn.referenceBase):
currentAlignmentColumn.matchCount += 1
else:
currentAlignmentColumn.mismatchCount += 1
# Count the nucleotide
if (currentBase == 'A'):
currentAlignmentColumn.aCount += 1
elif (currentBase == 'G'):
currentAlignmentColumn.gCount += 1
elif (currentBase == 'C'):
currentAlignmentColumn.cCount += 1
elif (currentBase == 'T'):
currentAlignmentColumn.tCount += 1
else:
print('Unknown Base found in Alignment at position ' + str(currentAlignmentColumn.referencePosition) + ':' + currentBase)
raise Exception('Unknown Base in Alignment')
# TODO: What if the query insertion sequence is longer than one base?
# Maybe I can only adjust one base per iteration, is that okay? Probably for the Best, actually..
# Don't worry bout it for now.
# Calculate highest frequency base
# I hope this algorithm makes sense, probably there is a smarter way to do it.
if(currentAlignmentColumn.aCount >= currentAlignmentColumn.gCount and currentAlignmentColumn.aCount >= currentAlignmentColumn.cCount and currentAlignmentColumn.aCount >= currentAlignmentColumn.tCount):
mostFrequentBase = 'A'
mostFrequentBaseCount = currentAlignmentColumn.aCount
elif(currentAlignmentColumn.gCount >= currentAlignmentColumn.cCount and currentAlignmentColumn.gCount >= currentAlignmentColumn.tCount):
mostFrequentBase = 'G'
mostFrequentBaseCount = currentAlignmentColumn.gCount
elif(currentAlignmentColumn.cCount >= currentAlignmentColumn.tCount):
mostFrequentBase = 'C'
mostFrequentBaseCount = currentAlignmentColumn.cCount
else:
mostFrequentBase = 'T'
mostFrequentBaseCount = currentAlignmentColumn.tCount
# Add the next base to the new consensus sequence
if (currentAlignmentColumn.matchCount >= currentAlignmentColumn.mismatchCount and currentAlignmentColumn.matchCount >= currentAlignmentColumn.inCount and currentAlignmentColumn.matchCount >= currentAlignmentColumn.delCount):
# Aligned bases match the reference, add reference base to the consensus.
referenceAdjustment='-'
newConsensusSequence += currentAlignmentColumn.referenceBase
elif (currentAlignmentColumn.inCount >= currentAlignmentColumn.mismatchCount and currentAlignmentColumn.inCount >= currentAlignmentColumn.delCount):
# Aligned bases show an insertion.
# Add the Reference Base and the Insertion Base to the consensus.
totalSequenceAdjustments += 1
referenceAdjustment='I'
newConsensusSequence += currentAlignmentColumn.referenceBase + mostFrequentBase
self.wranglerLog.write(str(currentAlignmentColumn.referencePosition) + ':Insertion' +
'\n(' + str(currentAlignmentColumn.inCount) + '/' + str(currentAlignmentColumn.alignedCount) + ') = ' + str((100.0 * currentAlignmentColumn.inCount) / currentAlignmentColumn.alignedCount) + '% of aligned reads'
'\n(' + currentAlignmentColumn.referenceBase + ' > ' + currentAlignmentColumn.referenceBase + mostFrequentBase + ')' +
'\n')
#TODO: I need to insert multiple bases, if that is waht the alignment suggests.
elif (currentAlignmentColumn.delCount >= currentAlignmentColumn.mismatchCount):
# Reads show a deletion.
# Don't add anything to the consensus.
totalSequenceAdjustments += 1
referenceAdjustment='D'
self.wranglerLog.write(str(currentAlignmentColumn.referencePosition) + ':Deletion' +
'\n(' + str(currentAlignmentColumn.delCount) + '/' + str(currentAlignmentColumn.alignedCount) + ') = ' + str((100.0 * currentAlignmentColumn.delCount) / currentAlignmentColumn.alignedCount) + '% of aligned reads'
'\n(' + currentAlignmentColumn.referenceBase + ' > _)' +
'\n')
else:
# Mismatch base.
# Add the highest read count base to the reference.
# It might actually be the same base as the reference,
# Because this just means there are more mismatches than matches.
# Problematic base, at least we'll notice here.
# TODO: What to do with highly heterozygous Positions?
# I should report those that look particularly heterozygous, somewhere.
newConsensusSequence += mostFrequentBase
totalSequenceAdjustments += 1
referenceAdjustment='M'
self.wranglerLog.write(str(currentAlignmentColumn.referencePosition) + ':Mismatch' +
'\n(' + str(mostFrequentBaseCount) + '/' + str(currentAlignmentColumn.alignedCount) + ') = ' + str((100.0 * mostFrequentBaseCount) / currentAlignmentColumn.alignedCount) + '% of aligned reads'
'\n(' + currentAlignmentColumn.referenceBase + ' > ' + mostFrequentBase + ')' +
'\n')
# Write a line to the alignment Summary
alignmentSummaryFile.write(str(currentAlignmentColumn.referencePosition) +
',' + str(currentAlignmentColumn.referenceBase) +
',' + str(referenceAdjustment) +
',' + str(currentAlignmentColumn.alignedCount) +
',' + str(currentAlignmentColumn.unalignedCount) +
',' + str(currentAlignmentColumn.matchCount) +
',' + str(currentAlignmentColumn.mismatchCount) +
',' + str(currentAlignmentColumn.inCount) +
',' + str(currentAlignmentColumn.delCount) +
',' + str(currentAlignmentColumn.aCount) +
',' + str(currentAlignmentColumn.gCount) +
',' + str(currentAlignmentColumn.cCount) +
',' + str(currentAlignmentColumn.tCount) +
'\n')
alignmentInfo.alignmentColumns.append(currentAlignmentColumn)
print('\nTotal Sequence Adjustments:' + str(totalSequenceAdjustments) + ' (How many bases the consensus differs from the reference.)\n')
# Write the newly constructed consensus sequence.
currentConsensusSequenceFileName = join(alignmentOutputDirectory, 'Consensus.fasta')
consensusWriter = createOutputFile(currentConsensusSequenceFileName)
# TODO: How to i give this a better name? Can I find a gene guess or something?
sequenceID = "Consensus_Sequence"
write([SeqRecord(Seq(newConsensusSequence,
IUPAC.unambiguous_dna),
id=sequenceID, description="") ], consensusWriter, 'fasta')
consensusWriter.close()
self.wranglerLog.write('Total Sequence Adjustments:' + str(totalSequenceAdjustments) + '\n')
# Close Summary Files
alignmentSummaryFile.close()
#adjustedBasesSummaryFile.close()
return alignmentInfo
def phaseHeterozygousReads(self):
# TODO: Should this method accept a cluster count?
# That will break some things. What things?
# This method is only called from this file, in the summarizeAnalysis method.
print('Splitting reads by heterozygous positions')
# Get a list of reads for later.
parsedReads = list(parse(self.readInput, self.readInputFormat))
heterozygousConsensusDirectory = join(self.outputRootDirectory,'HeterozygousAlignment')
# Open the bam file
print ('opening final alignment_bamfile')
bamfile = AlignmentFile(join(heterozygousConsensusDirectory,'alignment.bam'), 'rb')
# Load up the Alignment Reference file, we'll need it.
alignmentReferenceFileName = join(heterozygousConsensusDirectory,'AlignmentReference.fasta')
alignmentRef = list(parse(alignmentReferenceFileName, 'fasta'))[0]
# get list of AlignedReads
print ('Making a list of Aligned Reads.')
readIDs = []
for read in parsedReads:
if not read.id in readIDs:
readIDs.append(read.id)
readIDs.sort()
# Heterozygous base list
heterozygousBasesSummaryFile = createOutputFile(join(heterozygousConsensusDirectory, 'HeterozygousBases.txt'))
heterozygousBasesSummaryFile.write('List of Heterozygous Bases (0-based):\n')
if (self.snps is not None and len(self.snps) > 0):
# A string of SNPs was passed in, I don't need to calculate them myself.
# TODO: I could write alignment stats here, like I do when i self-calculate the hetero positions.
# This is just a simple list of 0-based positions.
for snp in self.snps:
heterozygousBasesSummaryFile.write(str(snp) + '\n')
else:
# get list of Heterozygous Positions
# TODO: I suppose I don't need to align 100% of reads to determine heterozygosity.
# Maybe this would speed up if i use a smaller alignment, or stop the loop after X reads
print('Getting a list of Heterozygous Positions:')
self.snps = []
pileupIterator = bamfile.pileup(alignmentRef.id)
for pileupColumn in pileupIterator:
readCount = 0
matchCount = 0
mismatchCount = 0
insCount = 0
delCount = 0
# dictionary of base counts.
referenceBase = alignmentRef.seq[pileupColumn.pos].upper()
# Iterate the Reads at this position. Each read at each position is either:
# ins, Del, match, mismatch.
#TODO: is it possible to exclude secondary/supplemetnary in the pileups method? No.
for pileupRead in pileupColumn.pileups:
#TODO: Important. Filter secondary / supplementary reads. This is causing problems, these secondary reads are FULL of snps.
# Difficulty: these parameters are on an aligned segment.
alignedSegmentObject = pileupRead.alignment
if(False):
pass
elif(alignedSegmentObject.is_secondary):
#print ('Secondary read at Position ' + str(pileupColumn.pos))
pass
elif(alignedSegmentObject.is_supplementary):
#print ('Supplementary read at Position ' + str(pileupColumn.pos))
pass
# Just trying some things, not sure what these mean.
#elif (alignedSegmentObject.is_unmapped):
# print('UNMAPPED READ!!!!!!!!!!!!!!!!!!! what does that mean?')
#elif (alignedSegmentObject.is_qcfail):
# print('This read was a QC failure. What does that mean?????????????')
else:
readCount += 1
# indels
if(pileupRead.is_del == 1):
delCount += 1
elif(pileupRead.indel > 0):
insCount += 1
else:
currentBase = pileupRead.alignment.query_sequence[pileupRead.query_position].upper()
if(currentBase == referenceBase):
matchCount += 1
else:
mismatchCount += 1
# This is a cheap way to stop analysis early. I will only analyze the first 250 reads.
# Potential problem: are these reads sorted somehow? Maybe my numbers are biased by only looking at the
# first reads
# Todo: This is another parameter that can be tuned. Add to inputs? Maybe.
maxAnalyzedReadCounts = 1000
if(readCount > maxAnalyzedReadCounts):
break
matchProportion = (1.0 * matchCount / readCount)
insertionProportion = (1.0 * insCount / readCount)
deletionProportion = (1.0 * delCount / readCount)
mismatchProportion = (1.0 * mismatchCount / readCount)
#print ('Position ' + str(pileupColumn.pos) + ', Coverage ' + str(pileupColumn.n) + ', Match/Mismatch : ' + str(matchCount) + '/' + str(mismatchCount))
#print ('Match Percentage ' + str(matchProportion))
# TODO: Should accepted match proprtion be a commandline parameter?
# if > 75% of bases match, this is not a heterzygous position
baseCutoff = .70
if(matchProportion > baseCutoff or insertionProportion > baseCutoff or deletionProportion > baseCutoff):
pass
#print ('Position ' + str(pileupColumn.pos) + ', Coverage ' + str(pileupColumn.n) + ', Deletion/Insertion/Match/Mismatch : ' + str(delCount) + '/' + str(insCount) + '/' + str(matchCount) + '/' + str(mismatchCount))
#print ('This position does not look heterozygous.')
# If coverage is very low, we should not use this position
# This logic is flawed, i think this is never working.
elif ((1.0 * pileupColumn.n / readCount) < .25):
pass
elif (mismatchProportion > baseCutoff):
pass
# These are the hardcoded values I used for the DRA analysis. Cheating.
# # I want to write a condition where we don't use the position if it's not clearly polymorphic.
# #elif (False):
# # pass
# # If the mismatch proportion is too high, what happens? What if there are 2 different bases that are mismatched, like if both my alleles have a different snp from reference. I'll miss that right now.
#
# # TEMP, this is very temporary. This is specific to a reference.
# # TODO : Fix these hard coded values.
# TODO: I don't really need this code, this is to ignore regions of my DRA reference.
# Instead, I can pass in a list of 1-based polymorphic positions to sort based on those. A "whitelist" instead of a "blacklist"
# # In a perfect world....I could tell what positions are heterozygous, but I can't.
# # I can tell if this sequence is a homopolymer though, but looking at the bases around it.....But that's not the correct thing to do.
# # I can keep this logic but make it a parameter. Big deletion regions are hard to analyze so I'm just ignoring them for now.
# elif(5890 <= pileupColumn.pos <= 5970):
# print('WARNING: I am skipping analysis on a region using hardcoded values, check this in allele_wrangler.')
# pass
# elif (6203 <= pileupColumn.pos <= 6212):
# print('WARNING: I am skipping analysis on a region using hardcoded values, check this in allele_wrangler.')
# pass
# # Big String of A's
# elif (774 <= pileupColumn.pos <= 796):
# print('WARNING: I am skipping analysis on a region using hardcoded values, check this in allele_wrangler.')
# pass
# #Known homopolymer positions....this is terrible programming.
# # I could at least pass these in ad ignored positions....
# elif (pileupColumn.pos in (403,430, 1479, 1510, 1683,
# 1991, 1996, 1997, 2003, 2009, 2093, 2100, 2133, 2134, 2191,
# 2262, 2289, 2294, 2342, 2449, 2450, 2524, 2647, 2663, 2732,
# 2895, 2902, 3113, 3114, 3180, 3197, 3362, 3396, 3453, 3542,
# 3551, 3665, 3832, 3903, 3953, 4108, 4109, 4400, 4639, 4698,
# 4703, 4769, 4785, 4786, 4828, 4878, 5084, 5301, 5302, 5449,
# 5575, 5597, 6155, 6279, 6280, 6314, 6375, 6376, 6712, 6755,
# 6790, 7084, 7631, 7718, 7769, 7971, 7978, 8132, 8133, 8134,
# 8314, 8315, 8352, 8476, 8477, 8478, 8642, 8650, 8651, 8652,
# 8653, 8654, 8655, 8656, 8657, 8698, 8725, 8753, 8759
# )):
# print('WARNING: I am skipping analysis on a region using hardcoded values, check this in allele_wrangler.')
# pass
else:
#heterozygousBasesSummaryFile.write (str(pileupColumn.pos) + ', Coverage ' + str(pileupColumn.n) + ', Deletion/Insertion/Match/Mismatch : ' + str(delCount) + '/' + str(insCount) + '/' + str(matchCount) + '/' + str(mismatchCount) + '\n')
heterozygousBasesSummaryFile.write(str(pileupColumn.pos) + ', Coverage ' + str(
pileupColumn.n) + ', Deletion/Insertion/Match/Mismatch : ' + str(delCount) + '/' + str(
insCount) + '/' + str(matchCount) + '/' + str(mismatchCount)
+ ' : ' + str(round(deletionProportion,2)) + '/'
+ str(round(insertionProportion, 2)) + '/'
+ str(round(matchProportion, 2)) + '/'
+ str(round(mismatchProportion, 2))
+ '\n')
self.snps.append(pileupColumn.pos)
heterozygousBasesSummaryFile.close()
#print ('Pileup Column # ' + str(pileupIterator))
print('Calculating read distance arrays:')
# I'm making this distance array. In this array, a 0 represents a Match. a 1 represents indels or substitutions.
# This way I can calculate "distance" in an arbitrary number of dimensions
# Distance is a euclidian way to represent how far away a read is from the consensus,
# based on the heterozygous positions. Each heterozygous position is a "dimension" in this space
distanceArrays = {}
for readID in readIDs:
# TODO: A Bug! Initializing this list as 0s will bias the results.
# TODO: Pileupcolumn loop is not hitting each read. Only...half sometimes. Some reads are not analyzed.
# Why? SPOTTED IT! bamfile.pileup has a default to maximum read depth of 8000
#distanceArrays[readID] = list([999] * len(self.snps))
distanceArrays[readID] = list([0] * len(self.snps))
# I spotted the bug!!! pileup defaults to maximum 8000 read depth. That's bad!.
pileupIterator = bamfile.pileup(alignmentRef.id,max_depth=99999999)
#pileupIterator = bamfile.pileup(alignmentRef.id)
for pileupColumn in pileupIterator:
currentColumn = pileupColumn.pos
# Only do this if the column number exists in our list of heterozygous positions
if currentColumn in self.snps:
heterozygousPositionIndex = self.snps.index(currentColumn)
currentAnalyzedReadCount = 0 # A debugging variable, i dont think I actually use this count.
referenceBase = alignmentRef.seq[currentColumn].upper()
for pileupRead in pileupColumn.pileups:
currentAnalyzedReadCount += 1
readID = pileupRead.alignment.query_name
#print('Pos:' + str(currentColumn) + ', Refbase:' + str(referenceBase) + ', Read:' + str(readID))
# In this model, the distance is either 0 or 1. This was intentional but
# Maybe we can tune the algorithm using these distances.
# This could actually be tuned to do the heterozygous split using ONLY snps.
# TODO: if we're having problems splitting based on homopolymers check this spot.
# Maybe, I want to count indels as 0, no distance.
# TODO: Something to try: indels are -1. SNPS are 1. Match = 0
# Maybe that would help the sorting?
# TODO: Newest idea. Default to 0. 1 is match, -1 is indels. -1 is mismatches. I think that's it.
if(pileupRead.is_del == 1):
distanceArrays[readID][heterozygousPositionIndex] = -1
elif(pileupRead.indel > 0):
distanceArrays[readID][heterozygousPositionIndex] = -1
else:
currentBase = pileupRead.alignment.query_sequence[pileupRead.query_position].upper()
if(currentBase == referenceBase):
#print('Assinging Match. Column=' + str(currentColumn) + ', CurrentBase:' + str(currentBase) + ', HeterozygousPosIndex=' + str(heterozygousPositionIndex))
distanceArrays[readID][heterozygousPositionIndex] = 1
else:
distanceArrays[readID][heterozygousPositionIndex] = -1
print('At position ' + str(heterozygousPositionIndex + 1) + ' I analyzed ' + str(currentAnalyzedReadCount) + ' reads.')
self.printDistanceArrays(distanceArrays, join(self.heterozygousDirectory, 'DistanceArrays.csv'))
# TODO: Im making 3 clusters. that worked. I need to make a parameter for cluster count.
clusteredReadIDs = self.clusterReads(distanceArrays, 2)
# Dictionary of results to return. Key is location of the consensus sequence.
# Value is the # of reads represented in this consensus alignment.
coverageResults = {}
for zeroBasedClusterIndex, readCluster in enumerate(clusteredReadIDs):
# I want to call the Strand (1 and 2), not Strand (0 and 1).
clusterIndex = zeroBasedClusterIndex + 1
clusteredReadIDs = readCluster.keys()
clusterOutputDir = join(self.outputRootDirectory, 'Strand' + str(clusterIndex) + 'ClusteredReads')
distanceArrayFileName = join(clusterOutputDir, 'Strand' + str(clusterIndex) + 'DistanceArrays.csv')
self.printDistanceArrays(readCluster, distanceArrayFileName)
readOutputFileName = join(clusterOutputDir, 'Strand' + str(clusterIndex) + 'Reads.' + self.readInputFormat)
readOutputFile = createOutputFile(readOutputFileName)
# Loop parsed reads, grab reads belonging to this cluster.
# FYI it looks like each input is clustered in the output, i haven't found a missing read yet. I should still check.
for readObject in parsedReads:
#print ('ReadClusterIndex=' + str(zeroBasedClusterIndex))
#print ('AllReadID=' + str(readObject.id))
for clusteredReadID in clusteredReadIDs:
#print ('clusteredReadID=' + str(clusteredReadID))
if (readObject.id == clusteredReadID):
write([readObject], readOutputFile, self.readInputFormat)
break
readOutputFile.close()
currentWranglerObject = AlleleWrangler(
readOutputFileName
, join(self.outputRootDirectory, 'Strand' + str(clusterIndex) + 'Alignment')
, join(self.heterozygousDirectory, 'AlignmentReference.fasta')
, 6
, self.numberThreads
, False
, self.snps)
currentCoverageResults = currentWranglerObject.analyzeReads()
# Merge the dictionaries of coverage values ane return them.
for key in currentCoverageResults.keys():
coverageResults[key] = currentCoverageResults[key]
print ('Done Phasing Reads.')
return coverageResults
