def calculate_euchromatin_ratios(kmer_generator):
euchromatin_ratio = {}
euchromatin_abundance = {
kmer: abundance
for kmer, abundance in kmer_generator.kmer_counters[False].items()
}
for kmer in euchromatin_abundance:
euchromatin_abundance[
kmer] = euchromatin_abundance[kmer] + euchromatin_abundance.get(
reverse_complement(kmer), 0)
heterochromatin_abundance = {
kmer: abundance
for kmer, abundance in kmer_generator.kmer_counters[True].items()
}
for kmer in heterochromatin_abundance:
heterochromatin_abundance[kmer] = heterochromatin_abundance[
kmer] + heterochromatin_abundance.get(reverse_complement(kmer), 0)
for kmer in euchromatin_abundance:
if kmer not in heterochromatin_abundance:
euchromatin_ratio[kmer] = 1
else:
euchromatin_ratio[kmer] = float(euchromatin_abundance[kmer] /
(heterochromatin_abundance[kmer] +
euchromatin_abundance[kmer]))
return euchromatin_ratio
def _get_pcr_products(fwd_primers_locations, rev_primers_locations,
max_pcr_product_length):
fwd_ranges, fwr_primers = kmer_locations_to_ranges(
fwd_primers_locations,
reverse=False,
extend_range=max_pcr_product_length)
rev_ranges, rev_primers = kmer_locations_to_ranges(
rev_primers_locations,
reverse=True,
extend_range=max_pcr_product_length)
# print(fwd_ranges)
pcr_products = []
intersection = fwd_ranges.intersect(rev_ranges, strandedness="opposite")
# print(intersection)
for chrom in intersection.chromosomes:
df = intersection[chrom].df
for index in range(df.shape[0]):
line = df.iloc[index]
try:
frw_primer = fwr_primers[(line.Chromosome.encode(),
line.Start)]
rev_primer = rev_primers[(line.Chromosome.encode(), line.End)]
except KeyError:
continue
if frw_primer.strand != rev_primer.strand and frw_primer.seq != reverse_complement(
rev_primer.seq):
pcr_products.append((frw_primer, rev_primer))
return pcr_products
def _get_pcr_products_old(fwd_primers_locations, rev_primers_locations,
max_pcr_product_length):
fwd_locs = [
PrimerAndLocation(0, loc.seq, loc.chrom_location,
loc.is_heterochromatic)
for loc in fwd_primers_locations
]
locs = fwd_locs
rev_locs = [
PrimerAndLocation(1, loc.seq, loc.chrom_location,
loc.is_heterochromatic)
for loc in rev_primers_locations
]
locs.extend(rev_locs)
locs = sorted(locs,
key=lambda PrimerAndLocation:
(PrimerAndLocation.chrom_location))
# pcr_products = [(loc1, loc2) for loc1, loc2 in zip(locs[:-1], locs[1:]) if (loc1.strand == 0) and (loc1.strand != loc2.strand) and (loc1.chrom_location[0] == loc2.chrom_location[0]) and abs(loc1.chrom_location[1] - loc2.chrom_location[1]) < max_pcr_product_length]
pcr_products = []
for loc1, loc2 in zip(locs[:-1], locs[1:]):
if ((loc1.strand == 0) and (loc1.strand != loc2.strand)
and (loc1.chrom_location[0] == loc2.chrom_location[0])
and (loc1.seq != reverse_complement(loc2.seq))
and (abs(loc1.chrom_location[1] - loc2.chrom_location[1]) <
max_pcr_product_length)):
pcr_products.append((loc1, loc2))
return pcr_products
def get_total_nondimer_pcr_products(pcr_products):
non_dimer_pcr_products = []
for product in pcr_products:
fwd_primer = product[0].seq
rev_primer = product[1].seq
if fwd_primer == reverse_complement(rev_primer):
continue
else:
non_dimer_pcr_products.append(product)
return non_dimer_pcr_products
def filter_out_kmers_by_revcomp(combination, sorted_primer_locations):
fwd_primer = combination[0]
rev_primer = combination[1]
if fwd_primer == reverse_complement(rev_primer):
num_products_fwd = len(sorted_primer_locations[fwd_primer])
num_products_rev = len(sorted_primer_locations[rev_primer])
if num_products_fwd < num_products_rev:
return fwd_primer
else:
return rev_primer
def filter_kmers_by_revcomp(kmers, kmer_counters, count_heterochromatic=False):
filt_out_kmers = []
for combination in combinations(kmers, 2):
fwd_primer = combination[0]
rev_primer = combination[1]
if fwd_primer.decode() == reverse_complement(rev_primer).decode():
num_products_fwd = kmer_counters[False][fwd_primer]
num_products_rev = kmer_counters[False][rev_primer]
if count_heterochromatic:
num_products_fwd += kmer_counters[True][fwd_primer]
num_products_rev += kmer_counters[True][rev_primer]
if num_products_fwd < num_products_rev:
filt_out_kmers.append(fwd_primer)
else:
filt_out_kmers.append(rev_primer)
kmers = [kmer for kmer in kmers if kmer not in filt_out_kmers]
return kmers
def main():
args = get_args()
genome_fhand = args['genome_fhand']
heterochromatic_regions_fhand = args['regions_fhand']
kmer_len = args['kmer_size']
cache_dir = Path(args['cache_dir'])
top_kmers = args['top_kmers']
num_sets = args['num_sets']
report_fhand = args['report_fhand']
forced_kmers = args['forced_kmers']
if not cache_dir.exists():
cache_dir.mkdir(exist_ok=True)
pcr_products_fhand = args['products_fhand']
kmers_to_keep = None
if forced_kmers:
forced_kmers = [kmer.encode() for kmer in forced_kmers]
kmers_to_keep = forced_kmers + [
reverse_complement(kmer) for kmer in forced_kmers
]
heterocromatic_fpath = heterochromatic_regions_fhand.name if heterochromatic_regions_fhand else None
kmers, kmers_locations = get_kmers(genome_fhand.name,
heterocromatic_fpath,
kmer_len,
cache_dir,
num_kmers_to_keep=top_kmers,
kmers_to_keep=kmers_to_keep)
if forced_kmers:
kmers = forced_kmers
primer_combinations = select_primers_combinations(
kmers, num_compatible_groups=num_sets)
product_results = get_pcr_products_in_sets(primer_combinations,
kmers,
kmers_locations,
max_pcr_product_length=10000)
pickle.dump(product_results, pcr_products_fhand, pickle.HIGHEST_PROTOCOL)
stats = get_stats_by_pair_in_sets(product_results,
kmers_locations=kmers_locations)
write_stats_in_excel(report_fhand.name, stats)
def get_pcr_products(kmer_locations, primers, max_pcr_product_length=5000):
pcr_products = {}
for combination in combinations(primers, 2):
fwd_primers = combination
fwd_primers_locations = []
rev_primers = [reverse_complement(primer) for primer in combination]
rev_primers_locations = []
for primer in fwd_primers:
fwd_primers_locations.extend(kmer_locations.get(primer, ""))
for primer in rev_primers:
rev_primers_locations.extend(kmer_locations.get(primer, ""))
pcr_products[combination] = _get_pcr_products(
fwd_primers_locations,
rev_primers_locations,
max_pcr_product_length=max_pcr_product_length)
return pcr_products
def kmer_locations_to_ranges(kmer_locations, reverse=False, extend_range=1):
chroms = []
starts = []
ends = []
strand = '-' if reverse else '+'
primer_strand = 1 if reverse else 0
seqs = []
is_hetero = []
indexed_primers = {}
for loc in kmer_locations:
chroms.append(loc.chrom_location[0].decode())
if reverse:
ends.append(loc.chrom_location[1])
start = loc.chrom_location[1] - extend_range
if start < 0:
start = 0
starts.append(start)
else:
starts.append(loc.chrom_location[1])
ends.append(loc.chrom_location[1] + extend_range)
seqs.append(reverse_complement(loc.seq) if reverse else loc.seq)
is_hetero = loc.is_heterochromatic
indexed_primers[(loc.chrom_location)] = PrimerAndLocation(
primer_strand, loc.seq, loc.chrom_location, loc.is_heterochromatic)
df = pd.DataFrame({
'Chromosome': chroms,
'Start': starts,
'End': ends,
'seq': seqs,
'Strand': strand,
'is_het': is_hetero
})
return pr.PyRanges(df), indexed_primers
def _get_union_sites_for_kmer(kmer, kmer_locations):
forward_unions = len(kmer_locations[kmer])
rev_kmer = reverse_complement(kmer)
rev_unions = len(kmer_locations[rev_kmer])
return forward_unions + rev_unions
def get_revcomp_locs(kmer, kmer_generator, get_heterochromatic_locs=False):
return kmer_generator.kmer_counters[False][reverse_complement(kmer)]