def read_or_make(cls, *, path_to_genome, path_to_index=None):
"""
Create an index for the genome and write to file.
Attempt to read from file instead if it already exists.
Default `path_to_index` appends the suffix ".index".
Returns the index.
RA, 2020-10-23
"""
from pathlib import Path
DEFAULT_SUFFIX = ".index"
path_to_genome = Path(path_to_genome)
path_to_index = Path(path_to_index
or (str(path_to_genome) + DEFAULT_SUFFIX))
assert path_to_genome.is_file()
if path_to_index.is_file():
return cls.read(path_to_index)
else:
from humdum.io import from_fasta
from humdum.utils import unlist1
return cls(unlist1(list(
from_fasta(path_to_genome))).seq).write(path_to_index)
def test_sw_on_data_small(self, verbose=0):
fa = Path(
__file__).parent / "data_for_tests/data_small/genome.chr22.5K.fa"
reference = str(unlist1(list(from_fasta(fa))).seq)
in_file = list((Path(__file__).parent /
"data_for_tests/data_small/").glob("*.sam")).pop()
max_reads = 2
for (read, __) in zip(from_sam(in_file), range(max_reads)):
read: Read
ref = reference
query = read.seq
aligner = SmithWaterman()
for alignment in aligner(ref=ref, query=query):
if verbose:
print(alignment.cigar, ' vs ', read.cigar)
print(read.mapq, ' vs ', alignment.score)
x, y, z = alignment.visualize(ref=ref, query=query)
print(x)
print(y)
print(z)
print(alignment.matching_subsegments(), ' vs ', read.cigar)
self.assertEqual(
alignment.cigar, read.cigar,
f'{alignment.cigar} is not equal to cigar from sam file {read.cigar}'
)
def test_on_data_big(self):
source_path = data_root / "data"
files = list(source_path.glob("*.fa.gz"))
assert files
for file in files:
for genome in from_fasta(file):
self.assertEqual(len(genome.seq), 51304566)
self.assertTrue(genome.seq.strip("N").endswith("CGGATT"))
def test_reads_well(self):
desc = "Hello"
seq1 = "ABC"
seq2 = "DEF"
with NamedTemporaryFile(mode='w') as fn:
print(*[">" + desc, seq1, seq2], sep='\n', file=fn, flush=True)
record = first(from_fasta(fn.name))
self.assertEqual(record.desc, desc)
self.assertEqual(record.seq, seq1 + seq2)
def test_data_small_vs_biopython(self):
source_path = data_root / "data_small"
files = list(source_path.glob("*.fa"))
assert files
from Bio import SeqIO
for file in files:
reference_reads = list(SeqIO.parse(file, format='fasta'))
candidate_reads = list(from_fasta(file))
self.assertEqual(len(reference_reads), len(candidate_reads))
for (reference, candidate) in zip(reference_reads,
candidate_reads):
self.assertIsInstance(reference, SeqIO.SeqRecord)
self.assertIsInstance(candidate, Sequence)
self.assertEqual(str(reference.seq), candidate.seq)
self.assertEqual(reference.description, candidate.desc)
def from_files(cls, *, fa, fq1, fq2):
"""
Reference genome file `fa`.
FASTQ files `fq1` and `fq2`.
Creates an instance of AllTheKingsHorses and
yields from its map_paired(...) member function.
"""
ref_genome = unlist1(from_fasta(fa))
index = GenomeIndex.read_or_make(path_to_genome=fa)
aligner = SequenceAligner()
atkh = AllTheKingsHorses(genome_index=index,
sequence_aligner=aligner,
ref_genome=ref_genome)
class _:
headers = atkh.headers()
alignments = atkh.map_paired(fq1, fq2)
return _
def test_fails_when_many(self):
with NamedTemporaryFile(mode='w') as fn:
print(*[">A", "N", ">B", "N"], sep='\n', file=fn, flush=True)
with self.assertRaises(AssertionError):
list(from_fasta(fn.name))
def read_or_make(cls, *, path_to_genome, ignored=None):
return cls(unlist1(list(from_fasta(path_to_genome))).seq)