def main(args):
# keep track of chromosome since this will be run with bash script
chrom = args['<chrom>']
# define output prefix
outprefix = args['<out>']
# make Fasta object for genome of choice, e.g. hg19
genome = Fasta(args['<fasta>'], as_raw=True)
cas_list = args['<cas_list>'].split(',')
# get set of positions for each type of cas
for cas in cas_list:
current_cas = cas_obj.get_cas_enzyme(
cas, os.path.join(cas_obj_path, 'CAS_LIST.txt'))
for_starts, rev_starts = find_spec_pams(current_cas,
str(genome[str(chrom)]),
orient=current_cas.primeness)
savestr_for = f'{outprefix}' + str(chrom) + '_' + str(
cas) + '_pam_sites_for.npy'
savestr_rev = f'{outprefix}' + str(chrom) + '_' + str(
cas) + '_pam_sites_rev.npy'
np.save(savestr_for, list(for_starts))
np.save(savestr_rev, list(rev_starts))
def get_made_broke_pams(df, chrom, ref_genome):
"""
Apply makes_breaks_pams to a df.
:param df: gens df generated by get_chr_tables.sh, available on EF github.
:param chrom: chromosome currently being analyzed.
:param ref_genome: ref_genome fasta, pyfaidx format.
:return: dataframe with indicators for whether each variant makes/breaks PAMs, pd df.
"""
FULL_CAS_LIST = cas_obj.get_cas_list(
os.path.join(cas_obj_path, "CAS_LIST.txt"))
for cas in cas_list:
if cas not in FULL_CAS_LIST:
logging.info(f"Skipping {cas}, not in CAS_LIST.txt")
continue
current_cas = cas_obj.get_cas_enzyme(
cas, os.path.join(cas_obj_path, "CAS_LIST.txt"))
makes, breaks = zip(*df.apply(
lambda row: makes_breaks_pam(current_cas, chrom, row["pos"], row[
"ref"], row["alt"], ref_genome),
axis=1,
))
df[f"makes_{cas}"] = makes
df[f"breaks_{cas}"] = breaks
return df
def adjusted_length(row):
"""
Adds on the length of the PAM to the sequnce length.
"""
cas = cas_object.get_cas_enzyme(row["cas_type"])
if row["strand"] == "positive":
return (row["start"], row["stop"] + len(cas.forwardPam))
else:
return (row["start"] - len(cas.forwardPam), row["stop"])
colors_formatted = []
for color in colors:
colors_formatted.append(
str(color[:3]).replace(' ', '').replace('[', '').replace(']', ''))
# map Cas proteins to colors
cas_to_colors = dict(zip(cas_list, colors_formatted))
bed_dfs = []
for chrom in chroms:
for cas in cas_list:
print(cas)
cas_info = cas_obj.get_cas_enzyme(
cas, os.path.join(cas_obj_path, 'CAS_LIST.txt'))
pam_size = len(cas_info.forwardPam)
for_pams = []
rev_pams = []
chroms_for = []
chroms_rev = []
with gzip.open(
f'/pollard/data/projects/AlleleAnalyzer_data/pam_sites_hg38/pam_sites_hg38_txt/chr{chrom}_{cas}_pam_sites_for.txt.gz',
'rb') as f:
for line in f.readlines():
for_pams.append(int(float(line.strip())) - 1)
chroms_for.append('chr' + str(chrom))
with gzip.open(
f'/pollard/data/projects/AlleleAnalyzer_data/pam_sites_hg38/pam_sites_hg38_txt/chr{chrom}_{cas}_pam_sites_rev.txt.gz',
'rb') as f:
for line in f.readlines():
def main(args):
logging.info(args)
out = args["<out>"]
pams_dir = args["<pams_dir>"]
gens = args["<gens_file>"]
guide_len = int(args["--guide_len"])
ref_genome = Fasta(args["<ref_genome_fasta>"], as_raw=True)
global cas_list
cas_list = list(args["<cas>"].split(","))
# Read in gens and chroms file, and see if gens file needs to be split.
gens = pd.read_hdf(gens, "all")
if gens.empty:
print('No variants in this region.')
exit()
chroms = dict(Counter(gens.chrom)).keys()
if len(chroms) > 1:
gens = split_gens(gens, list(chroms))
else:
gens = [gens]
fasta_chrom = list(ref_genome.keys())[0].startswith("chr")
chroms = [norm_chr(ch, fasta_chrom) for ch in list(chroms)]
# # Add check to make sure the correct FASTA file was loaded. - this is too glitchy
# if set(chroms) != set(list(ref_genome.keys())):
# logging.error(f"{args['<gens_file>']} chromosomes/notations differ from {args['<ref_genome_fasta>']}: {chroms} and {list(ref_genome.keys())}.")
# exit(1)
# save locations of PAM proximal variants to dictionary
pam_prox_vars = {}
# get variants within sgRNA region for 3 prime PAMs (20 bp upstream of for pos and vice versa)
FULL_CAS_LIST = cas_obj.get_cas_list(
os.path.join(cas_obj_path, "CAS_LIST.txt"))
for cas in cas_list:
if cas not in FULL_CAS_LIST:
logging.info(f"Skipping {cas}, not in CAS_LIST.txt")
cas_list.remove(cas)
combined_df = []
for i, chrom in enumerate(chroms):
chr_variants = set(gens[i]["pos"].tolist())
for cas in cas_list:
current_cas = cas_obj.get_cas_enzyme(
cas, os.path.join(cas_obj_path, "CAS_LIST.txt"))
logging.info(f"Evaluating {current_cas.name} at {chrom}.")
cas_prox_vars = []
pam_dict = {}
pam_for_pos = np.load(
os.path.join(pams_dir,
f"{chrom}_{cas}_pam_sites_for.npy")).tolist()
pam_rev_pos = np.load(
os.path.join(pams_dir,
f"{chrom}_{cas}_pam_sites_rev.npy")).tolist()
if current_cas.primeness == "3'":
for pos in pam_for_pos:
prox_vars = set(get_range_upstream(
pos, guide_len)) & chr_variants
cas_prox_vars.extend(prox_vars)
pam_dict[pos] = prox_vars
for pos in pam_rev_pos:
prox_vars = set(get_range_downstream(
pos, guide_len)) & chr_variants
cas_prox_vars.extend(prox_vars)
pam_dict[pos] = prox_vars
elif current_cas.primeness == "5'":
for pos in pam_for_pos:
prox_vars = set(get_range_downstream(
pos, guide_len)) & chr_variants
cas_prox_vars.extend(prox_vars)
pam_dict[pos] = prox_vars
for pos in pam_rev_pos:
prox_vars = set(get_range_upstream(
pos, guide_len)) & chr_variants
cas_prox_vars.extend(prox_vars)
pam_dict[pos] = prox_vars
pam_prox_vars[cas] = cas_prox_vars
chrdf = get_made_broke_pams(gens[i], chrom, ref_genome)
for cas in cas_list:
# print(cas)
spec_pam_prox_vars = pam_prox_vars[cas]
chrdf[f"var_near_{cas}"] = chrdf["pos"].isin(spec_pam_prox_vars)
cas_cols = []
for cas in cas_list:
prelim_cols = [
w.replace("cas", cas)
for w in ["makes_cas", "breaks_cas", "var_near_cas"]
]
cas_cols.extend(prelim_cols)
keepcols = ["chrom", "pos", "ref", "alt"] + cas_cols
chrdf = chrdf[keepcols]
combined_df.append(chrdf)
combined_df = pd.concat(combined_df)
combined_df.to_hdf(
f"{out}.h5",
"all",
mode="w",
format="table",
data_columns=True,
complib="blosc",
)
add_metadata(f"{out}.h5", args, os.path.basename(__file__), __version__,
"Annotation")
logging.info("Done.")