def __init__(self, format, inputPath): """ Constructs the bioinformatics file using the inputPath The input file will have these minimum requirements: - rsID - chromosome - position - sample genotype If input filetype is not specified, then send through parseNstream to predict filetype? """ if os.path.exists(inputPath): self.inputPath = inputPath self.referencePath = '/Volumes/teyden/BIOINFORMATICS/23andme2vcf/23andme_v4_hg19_ref.txt' self.genome = parseNstream._23andmeObject(self.inputPath, sortOption='CHROMOSOME') ## self.chrList = [ '%s' % i for i in range(1,26)] # + ['X', 'Y', 'MT']
133275309, 114364328, 107043718, 101991189, 90338345, 83257441, \
80373285, 58617616, 64444167, 46709983, 50818468, 156040895, 57227415, 16569]
self.positions = []
self.minSNP = 0
self.maxSNP = 0
def size(self):
return self.chrLengths[self.num-1]
def totalSNPs(self):
return len(self.positions)
ChromosomeDict = _23andmeObject("../../BIO_DATA/raw23andme_SNPs_teyden.txt", 'CHROMOSOME')
ChromosomePositions = {}
Chromosomes = {} # Dictionary of Chromosome objects
### (1) One list of SNP positions for every chromosome (1-22, X, Y, MT)
"""
ChromosomePositions['1'] = a list of positions for chromosome 1
... etc.
"""
# Make dict of all SNP positions. Keys = Chromosome #, Values = lst of positions
for chrom in ChromosomeDict:
ChromosomePositions[chrom] = ChromosomeDict[chrom].keys()
ChromosomePositions[chrom].sort() # Could go without
Chromosomes[chrom] = Chromosome(chrom=chrom)
Chromosomes[chrom].positions = ChromosomePositions[chrom]
def concatenateSNPs(RefSNPFile, UserSNPFile):
"""
Structure of RefSNPdict and UserSNPdict:
key = chromosome string as '1', '2', ..., '23' (X), '24' (Y), '25', (MT/M)
(*) Reference allele is only ONE letter.
- Genotype is a string of len=2, the original allele representations from the 23andme file
- Variant(s) represents any letters different from the Reference, if none then given '-'
- Match Score (0,1,2): the SNP is given a score of ...
- 2 (homozygous) if both letters are the same as the Reference
- 1 (heterozygous) if one letter is the same as the Reference
- 0 (recessive) if neither match the reference
(*) Variants given a match score of 0 may need to be switched to the
opposite letters (A to T, C to G and vice versa); must check to
confirm which are minus vs. plus strands)
### BASIC FORMAT:
#chrNum position Ref Genotype Variant Matches(0,1,2)
1 1000 A AA -
RETURNS: none, streams to file...
"""
RefSNPdict = parseNstream._referenceObject(RefSNPFile, 'CHROMOSOME')
UserSNPdict = parseNstream._23andmeObject(UserSNPFile, 'CHROMOSOME')
nonMatchedRSIDs = []
numMatchedRSIDs = 0
snpsConcatenated = 0
indel = {}
F = open(DEFAULT_OUTPUT_FILEPATH, 'w')
for chr in parseNstream.CHROMOSOME_LIST:
positions = RefSNPdict[chr].keys()
positions.sort()
indel[chr] = {}
for pos in positions:
refAllele = RefSNPdict[chr][pos][GENOTYPE].upper()
rsid = RefSNPdict[chr][pos][RSID]
# Check allele cases for ref in the user's snps
if pos in UserSNPdict[chr]:
if rsid != UserSNPdict[chr][pos][RSID]:
nonMatchedRSIDs += [(rsid, UserSNPdict[chr][pos][RSID])]
if isIndel(UserSNPdict[chr][pos][GENOTYPE]):
indel[chr][pos] = {
RSID: UserSNPdict[chr][pos][RSID],
GENOTYPE: UserSNPdict[chr][pos][GENOTYPE]
}
else:
numMatchedRSIDs += 1
sampleAllele = UserSNPdict[chr][pos][GENOTYPE] # 3
# Chromosomes 1-22 will have an allele pair, hence str len of 2
if len(sampleAllele) == 2:
if sampleAllele[0] == '-' and sampleAllele[1] == '-':
variant = '-'
score = '-'
elif sampleAllele[0].lower() == refAllele.lower() and sampleAllele[1].lower() == refAllele.lower():
variant = '-'
score = 2
elif sampleAllele[0].lower() == refAllele.lower():
variant = sampleAllele[1]
score = 1
elif sampleAllele[1].lower() == refAllele.lower():
variant = sampleAllele[0]
score = 1
else:
variant = sampleAllele
score = 0
# Mitochondria, X, and Y chromosome alleles are of length 1
elif len(sampleAllele) == 1:
if sampleAllele == '-':
variant = '-'
score = '-'
elif sampleAllele.lower() == refAllele.lower():
variant = '-'
score = 1
else:
variant = sampleAllele
score = 0
if score != '-':
snpsConcatenated += 1
if variant == '-':
F.write('%s' % sampleAllele[0])
else:
F.write('%s' % variant)
F.close()
parseNstream.printMsg("Completed scoring user SNPs to REF SNPs. Streamed to a .FASTA file (path=%s)" % DEFAULT_OUTPUT_FILEPATH)
parseNstream.printMsg('Number of matched rsid values: %s' % numMatchedRSIDs)
print "Only the RSID's of the user that existed in the reference data base were used. Also, INDELs were not concatenated."
parseNstream.printMsg('Number of unmatched rsid values: %s' % len(nonMatchedRSIDs))
parseNstream.printMsg('Number of SNPs concatenated: %s' % snpsConcatenated)
def scoreAlleles(RefSNPFile, UserSNPFile):
"""
Structure of RefSNPdict and UserSNPdict:
key = chromosome string as '1', '2', ..., '23' (X), '24' (Y), '25', (MT/M)
(*) Reference allele is only ONE letter.
- Genotype is a string of len=2, the original allele representations from the 23andme file
- Variant(s) represents any letters different from the Reference, if none then given '-'
- Match Score (0,1,2): the SNP is given a score of ...
- 2 (homozygous) if both letters are the same as the Reference
- 1 (heterozygous) if one letter is the same as the Reference
- 0 (recessive) if neither match the reference
(*) Variants given a match score of 0 may need to be switched to the
opposite letters (A to T, C to G and vice versa); must check to
confirm which are minus vs. plus strands)
### BASIC FORMAT:
#chrNum position Ref Genotype Variant Matches(0,1,2)
1 1000 A AA -
RETURNS: none, streams to file...
"""
RefSNPdict = parseNstream._referenceObject(RefSNPFile, 'CHROMOSOME')
UserSNPdict = parseNstream._23andmeObject(UserSNPFile, 'CHROMOSOME')
# Keep track of RSIDs in the that do not match in position and rsid value in tuples (ref rsid, user rsid)
nonMatchedRSIDs = []
# Count number of RSIDs whose positions in a chr match for both ref and user, and report it at the end.
numMatchedRSIDs = 0
# Keep track of indels
indel = {}
F = open(DEFAULT_OUTPUT_FILEPATH, 'w')
F.write('#Chrom\tRSID\tPos\tRef\tGenotype\tVariant(s)\tMatch Score\n')
for chr in parseNstream.CHROMOSOME_LIST:
positions = RefSNPdict[chr].keys()
positions.sort()
indel[chr] = {}
for pos in positions:
refAllele = RefSNPdict[chr][pos][GENOTYPE].upper()
rsid = RefSNPdict[chr][pos][RSID]
# Check allele cases for ref in the user's snps
if pos in UserSNPdict[chr]:
if rsid != UserSNPdict[chr][pos][RSID]:
nonMatchedRSIDs += [(rsid, UserSNPdict[chr][pos][RSID])]
if isIndel(UserSNPdict[chr][pos][GENOTYPE]):
indel[chr][pos] = {
RSID: UserSNPdict[chr][pos][RSID],
GENOTYPE: UserSNPdict[chr][pos][GENOTYPE]
}
else:
numMatchedRSIDs += 1
sampleAllele = UserSNPdict[chr][pos][GENOTYPE] # 3
# Chromosomes 1-22 will have an allele pair, hence str len of 2
if len(sampleAllele) == 2:
# No base call at the current rsid. [BLANK]
if sampleAllele[0] == '-' and sampleAllele[1] == '-':
variant = '-'
score = '-'
# Score 2, no variants. Both letters are homologous to the reference. [GRAY - default]
elif sampleAllele[0].lower() == refAllele.lower() and sampleAllele[1].lower() == refAllele.lower():
variant = '-'
score = 2
# Score 1, one variant. Second letter is homologous to the reference. [GREEN - success]
elif sampleAllele[0].lower() == refAllele.lower():
variant = sampleAllele[1]
score = 1
# Score 1, one variant. First letter is homologous to the reference. [GREEN - success]
elif sampleAllele[1].lower() == refAllele.lower():
variant = sampleAllele[0]
score = 1
# Score 0, two variants. No homology to the reference. [ORANGE - primary]
else:
variant = sampleAllele
score = 0
# Mitochondria, X, and Y chromosome alleles are of length 1
elif len(sampleAllele) == 1:
# No base call at the current rsid. [BLANK]
if sampleAllele == '-':
variant = '-'
score = '-'
# Score 1, no variants. Only letter is homologous to the reference. [GRAY - default]
elif sampleAllele.lower() == refAllele.lower():
variant = '-'
score = 1
else:
variant = sampleAllele
score = 0
F.write('%s\t%s\t%s\t%s\t%s\t%s\t%s\n' % (chr, rsid, pos, refAllele, sampleAllele, variant, score))
F.close()
parseNstream.printMsg("Completed scoring user SNPs to REF SNPs. Streamed to a .teyden (LOL) file format (path=%s)" % DEFAULT_OUTPUT_FILEPATH)
parseNstream.printMsg('Number of matched rsid values: %s' % numMatchedRSIDs)
parseNstream.printMsg('Number of unmatched rsid values: %s' % len(nonMatchedRSIDs))
return indel