def test_somatic_variant_with_2_supporting_rna_reads():
variant = Variant("14", 105849746, "G", "A", grch38)
base_dir = "data/somatic-variant-with-2-supporting-rna-reads/"
normal_reads = load_bam(base_dir +
"normal.14.105849746.G.A.no-alt.sorted.bam")
tumor_reads = load_bam(base_dir +
"tumor.14.105849746.G.A.many-alt.sorted.bam")
rna_reads = load_bam(base_dir + "rna.14.105849746.G.A.2-alt.sorted.bam")
read_creator = ReadCollector()
normal_sample_variant_reads = read_creator.allele_reads_supporting_variant(
variant=variant, alignment_file=normal_reads)
eq_(len(normal_sample_variant_reads), 0)
print(normal_sample_variant_reads)
tumor_sample_variant_reads = read_creator.allele_reads_supporting_variant(
variant=variant, alignment_file=tumor_reads)
print(tumor_sample_variant_reads)
eq_(len(tumor_sample_variant_reads), 8)
rna_sample_variant_reads = read_creator.allele_reads_supporting_variant(
variant=variant, alignment_file=rna_reads)
print(rna_sample_variant_reads)
eq_(len(rna_sample_variant_reads), 2)
# Arun went through the hassle of pulling out the exact read names
# in IGV
expected_variant_rna_read_names = {
"K00193:50:H5NKVBBXX:5:2202:6421:24964",
"K00193:50:H5NKVBBXX:5:2119:30908:1138",
}
for variant_read in rna_sample_variant_reads:
assert variant_read.name in expected_variant_rna_read_names
def test_somatic_variant_with_2_supporting_rna_reads():
variant = Variant("14", 105849746, "G", "A")
base_dir = "data/somatic-variant-with-2-supporting-rna-reads/"
normal_reads = load_bam(base_dir + "normal.14.105849746.G.A.no-alt.sorted.bam")
tumor_reads = load_bam(base_dir + "tumor.14.105849746.G.A.many-alt.sorted.bam")
rna_reads = load_bam(base_dir + "rna.14.105849746.G.A.2-alt.sorted.bam")
normal_sample_variant_reads = reads_supporting_variant(
variant=variant,
samfile=normal_reads)
eq_(len(normal_sample_variant_reads), 0)
print(normal_sample_variant_reads)
tumor_sample_variant_reads = reads_supporting_variant(
variant=variant,
samfile=tumor_reads)
print(tumor_sample_variant_reads)
eq_(len(tumor_sample_variant_reads), 8)
rna_sample_variant_reads = reads_supporting_variant(
variant=variant,
samfile=rna_reads)
print(rna_sample_variant_reads)
eq_(len(rna_sample_variant_reads), 2)
# Arun went through the hassle of pulling out the exact read names
# in IGV
expected_variant_rna_read_names = {
"K00193:50:H5NKVBBXX:5:2202:6421:24964",
"K00193:50:H5NKVBBXX:5:2119:30908:1138",
}
for variant_read in rna_sample_variant_reads:
assert variant_read.name in expected_variant_rna_read_names
def test_most_common_nucleotides_for_chr12_deletion():
samfile = load_bam("data/cancer-wgs-primary.chr12.bam")
chromosome = "chr12"
base1_location = 70091490
ref = "TTGTAGATGCTGCCTCTCC"
alt = ""
variant = Variant(
chromosome,
base1_location,
ref,
alt,
ensembl=ensembl_grch38)
variant_reads = reads_supporting_variant(
samfile=samfile,
chromosome=chromosome,
variant=variant)
consensus_sequence, chosen_counts, other_counts = most_common_nucleotides(
variant_reads)
print(chosen_counts)
print(other_counts)
eq_(len(chosen_counts), len(consensus_sequence))
eq_(len(other_counts), len(consensus_sequence))
assert other_counts.sum() < chosen_counts.sum(), \
"Counts for alternate nucleotides should not exceed the chosen sequence"
number_matching_reads = 0
for variant_read in variant_reads:
full_seq = variant_read.prefix + variant_read.allele + variant_read.suffix
number_matching_reads += (full_seq in consensus_sequence)
fraction_matching_reads = number_matching_reads / float(len(variant_reads))
print("Fraction matching reads is %d/%d = %f" % (
number_matching_reads, len(variant_reads), fraction_matching_reads))
assert fraction_matching_reads > 0.5, \
"Expected majority of reads to match consensus sequence"
def test_sequence_counts_snv():
samfile = load_bam("data/cancer-wgs-primary.chr12.bam")
chromosome = "chr12"
base1_location = 65857041
ref = "G"
alt = "C"
variant = Variant(chromosome, base1_location, ref, alt)
variant_reads = reads_supporting_variant(
samfile=samfile,
chromosome=chromosome,
variant=variant)
variant_sequences = reads_to_variant_sequences(
variant=variant,
reads=variant_reads,
preferred_sequence_length=61)
assert len(variant_sequences) == 1
for variant_sequence in variant_sequences:
print(variant_sequence)
eq_(variant_sequence.alt, alt)
eq_(len(variant_sequence.prefix), 30)
eq_(len(variant_sequence.suffix), 30)
eq_(
variant_sequence.prefix + variant_sequence.alt + variant_sequence.suffix,
variant_sequence.sequence)
def test_assemble_transcript_fragments_snv():
samfile = load_bam("data/cancer-wgs-primary.chr12.bam")
chromosome = "chr12"
base1_location = 65857041
ref = "G"
alt = "C"
variant = Variant(
contig=chromosome,
start=base1_location,
ref=ref,
alt=alt,
ensembl=ensembl_grch38)
variant_reads = reads_supporting_variant(
variant=variant,
samfile=samfile,
chromosome=chromosome,)
sequences = iterative_overlap_assembly(
initial_variant_sequences_from_reads(variant_reads),
min_overlap_size=30)
assert len(sequences) > 0
max_read_length = max(len(r) for r in variant_reads)
for s in sequences:
print("%s%s%s weight=%d length=%d" % (
s.prefix,
s.alt,
s.suffix,
len(s.reads),
len(s.sequence)))
eq_(s.alt, alt)
assert len(s) > max_read_length, \
"Expected assembled sequences to be longer than read length (%d)" % (
max_read_length,)
def test_group_unique_sequences():
samfile = load_bam("data/cancer-wgs-primary.chr12.bam")
chromosome = "chr12"
base1_location = 65857041
ref = "G"
alt = "C"
variant = Variant(
contig=chromosome,
start=base1_location,
ref=ref, alt=alt,
ensembl=ensembl_grch38)
variant_reads = reads_supporting_variant(
samfile=samfile,
chromosome=chromosome,
variant=variant)
print("%d variant reads: %s" % (
len(variant_reads), variant_reads))
groups = group_unique_sequences(
variant_reads,
max_prefix_size=30,
max_suffix_size=30)
print("%d unique sequences: %s" % (
len(groups), groups))
# there are some redundant reads, so we expect that the number of
# unique entries should be less than the total read partitions
assert len(variant_reads) > len(groups)
def variants_to_protein_sequences_dataframe(
expressed_vcf="data/b16.f10/b16.expressed.vcf",
not_expressed_vcf="data/b16.f10/b16.not-expressed.vcf",
tumor_rna_bam="data/b16.f10/b16.combined.sorted.bam",
min_mapping_quality=0,
max_protein_sequences_per_variant=1,
variant_sequence_assembly=False):
"""
Helper function to load pair of VCFs and tumor RNA BAM
and use them to generate a DataFrame of expressed variant protein
sequences.
"""
expressed_variants = load_vcf(expressed_vcf)
not_expressed_variants = load_vcf(not_expressed_vcf)
combined_variants = VariantCollection(
list(expressed_variants) + list(not_expressed_variants))
alignment_file = load_bam(tumor_rna_bam)
read_collector = ReadCollector(min_mapping_quality=min_mapping_quality)
read_evidence_gen = read_collector.read_evidence_generator(
variants=combined_variants, alignment_file=alignment_file)
creator = ProteinSequenceCreator(
max_protein_sequences_per_variant=max_protein_sequences_per_variant,
variant_sequence_assembly=variant_sequence_assembly)
protein_sequences_generator = \
creator.protein_sequences_from_read_evidence_generator(read_evidence_gen)
df = protein_sequences_generator_to_dataframe(protein_sequences_generator)
return df, expressed_variants, combined_variants
def variants_to_protein_sequences_dataframe(
expressed_vcf="data/b16.f10/b16.expressed.vcf",
not_expressed_vcf="data/b16.f10/b16.not-expressed.vcf",
tumor_rna_bam="data/b16.f10/b16.combined.sorted.bam",
min_mapping_quality=0,
max_protein_sequences_per_variant=1,
variant_sequence_assembly=False):
"""
Helper function to load pair of VCFs and tumor RNA BAM
and use them to generate a DataFrame of expressed variant protein
sequences.
"""
expressed_variants = load_vcf(expressed_vcf)
not_expressed_variants = load_vcf(not_expressed_vcf)
combined_variants = VariantCollection(
list(expressed_variants) + list(not_expressed_variants))
samfile = load_bam(tumor_rna_bam)
allele_reads_generator = reads_overlapping_variants(
variants=combined_variants,
samfile=samfile,
min_mapping_quality=min_mapping_quality)
protein_sequences_generator = reads_generator_to_protein_sequences_generator(
allele_reads_generator,
max_protein_sequences_per_variant=max_protein_sequences_per_variant,
variant_sequence_assembly=variant_sequence_assembly)
df = protein_sequences_generator_to_dataframe(protein_sequences_generator)
return df, expressed_variants, combined_variants
def test_assemble_transcript_fragments_snv():
samfile = load_bam("data/cancer-wgs-primary.chr12.bam")
chromosome = "chr12"
base1_location = 65857041
ref = "G"
alt = "C"
variant = Variant(contig=chromosome,
start=base1_location,
ref=ref,
alt=alt,
ensembl=ensembl_grch38)
variant_reads = reads_supporting_variant(
variant=variant,
samfile=samfile,
chromosome=chromosome,
)
sequences = iterative_overlap_assembly(
initial_variant_sequences_from_reads(variant_reads),
min_overlap_size=30)
assert len(sequences) > 0
max_read_length = max(len(r) for r in variant_reads)
for s in sequences:
print("%s%s%s weight=%d length=%d" %
(s.prefix, s.alt, s.suffix, len(s.reads), len(s.sequence)))
eq_(s.alt, alt)
if len(s.read_names) > 1:
# expect sequences supported by more than one read to be greater
# than the read length
assert len(s) > max_read_length, \
"Expected assembled sequences to be longer than read length (%d)" % (
max_read_length,)
def test_locus_reads_dataframe():
sam_all_variants = load_bam("data/b16.f10/b16.combined.bam")
n_reads_expected = 0
sam_path_single_variant = data_path(
"data/b16.f10/b16.f10.127a.aldh1b1.chr4.45802539.refG.altC.sam")
with open(sam_path_single_variant) as f:
for line in f:
if line.startswith("HWI"):
n_reads_expected += 1
# we know from inspecting the file that *one* of the reads overlapping this
# variant has a CIGAR string of N at the location before and thus we'll
# be missing that read.
#
# TODO: figure out what to do when the variant nucleotide is at the start or
# end of an exon, since that won't have mapping positions on both its left
# and right
n_reads_expected -= 1
print("Found %d sequences in %s" %
(n_reads_expected, sam_path_single_variant))
df = locus_reads_dataframe(samfile=sam_all_variants,
chromosome="chr4",
base1_position_before_variant=45802538,
base1_position_after_variant=45802540)
print(df)
eq_(len(df), n_reads_expected)
def test_locus_reads_dataframe():
sam_all_variants = load_bam("data/b16.f10/b16.combined.bam")
n_reads_expected = 0
sam_path_single_variant = data_path(
"data/b16.f10/b16.f10.127a.aldh1b1.chr4.45802539.refG.altC.sam")
with open(sam_path_single_variant) as f:
for line in f:
if line.startswith("HWI"):
n_reads_expected += 1
# we know from inspecting the file that *one* of the reads overlapping this
# variant has a CIGAR string of N at the location before and thus we'll
# be missing that read.
#
# TODO: figure out what to do when the variant nucleotide is at the start or
# end of an exon, since that won't have mapping positions on both its left
# and right
n_reads_expected -= 1
print("Found %d sequences in %s" % (n_reads_expected, sam_path_single_variant))
df = locus_reads_dataframe(
samfile=sam_all_variants,
chromosome="chr4",
base1_position_before_variant=45802538,
base1_position_after_variant=45802540)
print(df)
eq_(len(df), n_reads_expected)
def test_partition_variant_reads_deletion():
samfile = load_bam("data/cancer-wgs-primary.chr12.bam")
chromosome = "chr12"
base1_location = 70091490
ref = "TTGTAGATGCTGCCTCTCC"
alt = ""
variant = Variant(contig=chromosome, start=base1_location, ref=ref, alt=alt, ensembl=ensembl_grch38)
variant_reads = reads_supporting_variant(samfile=samfile, chromosome=chromosome, variant=variant)
assert len(variant_reads) > 1
for variant_read in variant_reads:
eq_(variant_read.allele, alt)
def test_somatic_variant_with_0_supporting_rna_reads():
variant = Variant("6", 90411765, "G", "A")
base_dir = "data/somatic-variant-with-0-supporting-rna-reads/"
normal_reads = load_bam(base_dir + "normal.6.90411765.G.A.sorted.bam")
tumor_reads = load_bam(base_dir + "tumor.6.90411765.G.A.sorted.bam")
rna_reads = load_bam(base_dir + "rna.6.90411765.G.A.sorted.bam")
normal_sample_variant_reads = reads_supporting_variant(
variant=variant, samfile=normal_reads)
eq_(len(normal_sample_variant_reads), 0)
print(normal_sample_variant_reads)
tumor_sample_variant_reads = reads_supporting_variant(variant=variant,
samfile=tumor_reads)
print(tumor_sample_variant_reads)
eq_(len(tumor_sample_variant_reads), 5)
rna_sample_variant_reads = reads_supporting_variant(variant=variant,
samfile=rna_reads)
print(rna_sample_variant_reads)
eq_(len(rna_sample_variant_reads), 0)
def test_translate_variant_collection():
variants = load_vcf("data/b16.f10/b16.vcf")
samfile = load_bam("data/b16.f10/b16.combined.sorted.bam")
result = list(translate_variants(reads_supporting_variants(variants, samfile)))
eq_(
len(result),
4,
"Expected %d translated variants but got %d: %s" % (
len(variants),
len(result),
result))
def test_translate_variant_collection():
variants = load_vcf("data/b16.f10/b16.vcf")
samfile = load_bam("data/b16.f10/b16.combined.sorted.bam")
read_evidence_gen = ReadCollector().read_evidence_generator(
variants,
samfile)
translation_gen = ProteinSequenceCreator().translate_variants(read_evidence_gen)
translations = list(translation_gen)
eq_(
len(translations),
4,
"Expected %d translated variants but got %d: %s" % (
len(variants),
len(translations),
translations))
def test_variants_to_protein_sequences_dataframe_filtered_all_reads_by_mapping_quality(
):
# since the B16 BAM has all MAPQ=255 values then all the reads should get dropped
# if we set the minimum quality to 256
variants = load_vcf("data/b16.f10/b16.vcf")
alignment_file = load_bam("data/b16.f10/b16.combined.sorted.bam")
read_collector = ReadCollector(min_mapping_quality=256)
read_evidence_gen = read_collector.read_evidence_generator(
variants=variants, alignment_file=alignment_file)
creator = ProteinSequenceCreator(max_protein_sequences_per_variant=1, )
protein_sequences_generator = creator.protein_sequences_from_read_evidence_generator(
read_evidence_gen)
df = protein_sequences_generator_to_dataframe(protein_sequences_generator)
print(df)
eq_(len(df), 0, "Expected 0 entries, got %d: %s" % (len(df), df))
def test_partition_variant_reads_deletion():
alignment_file = load_bam("data/cancer-wgs-primary.chr12.bam")
chromosome = "chr12"
base1_location = 70091490
ref = "TTGTAGATGCTGCCTCTCC"
alt = ""
variant = Variant(contig=chromosome,
start=base1_location,
ref=ref,
alt=alt,
ensembl=ensembl_grch38)
read_collector = ReadCollector()
read_evidence = read_collector.read_evidence_for_variant(
alignment_file=alignment_file, variant=variant)
assert len(read_evidence.alt_reads) > 1
for variant_read in read_evidence.alt_reads:
eq_(variant_read.allele, alt)
def test_partition_variant_reads_snv():
alignment_file = load_bam("data/cancer-wgs-primary.chr12.bam")
chromosome = "chr12"
base1_location = 65857041
ref = "G"
alt = "C"
variant = Variant(contig=chromosome,
start=base1_location,
ref=ref,
alt=alt,
ensembl=ensembl_grch38)
read_collector = ReadCollector()
read_evidence = read_collector.read_evidence_for_variant(
alignment_file=alignment_file, variant=variant)
alt_reads = read_evidence.alt_reads
assert len(alt_reads) > 1
for variant_read in alt_reads:
eq_(variant_read.allele, alt)
def test_variants_to_protein_sequences_dataframe_filtered_all_reads_by_mapping_quality():
# since the B16 BAM has all MAPQ=255 values then all the reads should get dropped
# if we set the minimum quality to 256
variants = load_vcf("data/b16.f10/b16.vcf")
samfile = load_bam("data/b16.f10/b16.combined.sorted.bam")
allele_reads_generator = reads_overlapping_variants(
variants=variants,
samfile=samfile,
min_mapping_quality=256)
protein_sequences_generator = reads_generator_to_protein_sequences_generator(
allele_reads_generator,
max_protein_sequences_per_variant=1)
df = protein_sequences_generator_to_dataframe(protein_sequences_generator)
print(df)
eq_(
len(df),
0,
"Expected 0 entries, got %d: %s" % (len(df), df))
def test_partition_variant_reads_deletion():
samfile = load_bam("data/cancer-wgs-primary.chr12.bam")
chromosome = "chr12"
base1_location = 70091490
ref = "TTGTAGATGCTGCCTCTCC"
alt = ""
variant = Variant(
contig=chromosome,
start=base1_location,
ref=ref,
alt=alt,
ensembl=ensembl_grch38)
variant_reads = reads_supporting_variant(
samfile=samfile,
chromosome=chromosome,
variant=variant)
assert len(variant_reads) > 1
for variant_read in variant_reads:
eq_(variant_read.allele, alt)
def test_protein_sequence_creator_protein_length():
variants = load_vcf("data/b16.f10/b16.vcf")
alignment_file = load_bam("data/b16.f10/b16.combined.sorted.bam")
read_collector = ReadCollector()
for desired_length in [21, 15, 10]:
creator = ProteinSequenceCreator(
max_protein_sequences_per_variant=1,
protein_sequence_length=desired_length)
read_evidence_gen = read_collector.read_evidence_generator(
variants=variants, alignment_file=alignment_file)
protein_sequences_generator = creator.protein_sequences_from_read_evidence_generator(
read_evidence_gen)
df = protein_sequences_generator_to_dataframe(
protein_sequences_generator)
print(df)
protein_sequences = df["amino_acids"]
print(protein_sequences)
protein_sequence_lengths = protein_sequences.str.len()
assert (protein_sequence_lengths == desired_length).all(), (
protein_sequence_lengths, )
def test_sequence_counts_snv():
samfile = load_bam("data/cancer-wgs-primary.chr12.bam")
chromosome = "chr12"
base1_location = 65857041
ref = "G"
alt = "C"
variant = Variant(chromosome, base1_location, ref, alt, grch38)
read_creator = ReadCollector()
variant_reads = read_creator.allele_reads_supporting_variant(
alignment_file=samfile, variant=variant)
variant_sequence_creator = VariantSequenceCreator(
preferred_sequence_length=61)
variant_sequences = variant_sequence_creator.reads_to_variant_sequences(
variant=variant, reads=variant_reads)
assert len(variant_sequences) == 1
for variant_sequence in variant_sequences:
print(variant_sequence)
eq_(variant_sequence.alt, alt)
eq_(len(variant_sequence.prefix), 30)
eq_(len(variant_sequence.suffix), 30)
eq_(
variant_sequence.prefix + variant_sequence.alt +
variant_sequence.suffix, variant_sequence.sequence)
def test_sequence_counts_snv():
samfile = load_bam("data/cancer-wgs-primary.chr12.bam")
chromosome = "chr12"
base1_location = 65857041
ref = "G"
alt = "C"
variant = Variant(chromosome, base1_location, ref, alt)
variant_reads = reads_supporting_variant(samfile=samfile,
chromosome=chromosome,
variant=variant)
variant_sequences = reads_to_variant_sequences(
variant=variant, reads=variant_reads, preferred_sequence_length=61)
assert len(variant_sequences) == 1
for variant_sequence in variant_sequences:
print(variant_sequence)
eq_(variant_sequence.alt, alt)
eq_(len(variant_sequence.prefix), 30)
eq_(len(variant_sequence.suffix), 30)
eq_(
variant_sequence.prefix + variant_sequence.alt +
variant_sequence.suffix, variant_sequence.sequence)
