def dseq2dguides(cfg):
"""
Wrapper around make guides function.
:param cfg: configuration dict.
"""
cfg['datad']=cfg[cfg['step']]
cfg['plotd']=cfg['datad']
dguideslinp=f"{cfg['datad']}/dguides.tsv"
dguides_nofltp=f"{cfg['datad']}/dguides_noflt.tsv"
dmutagenesisp=f"{cfg['datad']}/dmutagenesis.tsv"
dpam_strandsp=f"{cfg['datad']}/dpam_strands.csv"
if not exists(dguideslinp) or cfg['force']:
dmutagenesis=pd.read_csv(f"{cfg[cfg['step']-1]}/dmutagenesis.tsv",sep='\t')
if cfg['mutation_format']=='nucleotide':
dsequences=pd.read_csv(f"{cfg[cfg['step']-2]}/dsequences.tsv",sep='\t') #FIXME if numbering of steps is changed, this is gonna blow
dsequences,dmutagenesis=dinnucleotide2dsequencesproper(dsequences,dmutagenesis)
elif cfg['mutation_format']=='aminoacid':
dsequences=pd.read_csv(f"{cfg[cfg['step']-2]}/dsequences.tsv",sep='\t') #FIXME if numbering of steps is changed, this is gonna blow
if 'reverse_mutations' in cfg:
if cfg['reverse_mutations']:
# from beditor.lib.global_vars import stepi2cols
cols_dsequences=dsequences.columns.tolist()
dsequences=pd.merge(dsequences,
dmutagenesis,
how='inner',
left_on=['aminoacid: wild-type','codon: mutation','amino acid mutation'],
right_on=['amino acid','codon mutation','amino acid mutation'],
suffixes=['', ': dmutagenesis'])
dsequences['codon: wild-type']=dsequences['codon']
dsequences=dsequences.loc[:,cols_dsequences]
dsequences.to_csv(f"{cfg[cfg['step']]}/dsequences.tsv",sep='\t') #FIXME if numbering of steps is changed, this is gonna blow
# make pam table
dpam=pd.read_table(f'{dirname(realpath(__file__))}/../data/dpam.tsv')
if sum(dpam['PAM'].isin(cfg['pams']))!=len(cfg['pams']):
logging.error(f"PAM/s {cfg['pams']} are not supported {dpam['PAM'].tolist()}")
sys.exit(1)
dpam_strands=dpam2dpam_strands(dpam,pams=cfg['pams'])
dpam_strands.to_csv(dpam_strandsp,sep='\t')
if not (len(dsequences)==0 or len(dmutagenesis)==0):
dmutagenesis['strand']=dmutagenesis.apply(lambda x : x['mutation on strand'].replace(' strand',''),axis=1)
dmutagenesis.to_csv(dmutagenesisp,sep='\t')
dguideslin,dguides_noflt,err2idxs=make_guides(cfg,dsequences,
dmutagenesis,
dpam=dpam_strands,
test=cfg['test'],
# dbug=True,
)
if not dguides_noflt is None:
dguides_noflt.to_csv(dguides_nofltp,sep='\t')
if not ((dguideslin is None) and (err2idxs is None)):
dguideslin.to_csv(dguideslinp,sep='\t')
if cfg['test']:
logging.info(err2idxs)
with open(dguideslinp+'.err.json', 'w') as f:
json.dump(err2idxs, f)
else:
from beditor.lib.global_vars import saveemptytable
logging.warning('no guides designed; saving an empty table.')
saveemptytable(cfg,dguideslinp)
else:
from beditor.lib.global_vars import saveemptytable
logging.warning('no guides designed; saving an empty table.')
saveemptytable(cfg,dguideslinp)
import gc
gc.collect()
def get_seq_aminoacid(cfg, din):
"""
Fetches sequences if mutation format is amino acid
:param cfg: configuration dict
:param din: input data
:returns dsequences: dataframe with sequences
"""
import pyensembl
#import ensembl object that would fetch genes
# ensembl = pyensembl.EnsemblRelease(release=cfg['genomerelease'])
ensembl = pyensembl.EnsemblRelease(
species=pyensembl.species.Species.register(latin_name=cfg['host'],
synonyms=[cfg['host']],
reference_assemblies={
cfg['genomeassembly']:
(cfg['genomerelease'],
cfg['genomerelease']),
}),
release=cfg['genomerelease'])
din.index = range(len(din))
dbedp = '{}/dbedflank.bed'.format(cfg['datad'])
dbed = pd.DataFrame(columns=bed_colns)
terrpositions = []
terrnotfound = []
terrnoncoding = []
bedrowi = 0
# for i in trange(len(din)-1,desc='get positions for bedtools'):
for i in din.index:
if din.loc[i, 'transcript: id'] in ensembl.transcript_ids():
t = ensembl.transcript_by_id(din.loc[i, 'transcript: id'])
if t.is_protein_coding and t.contains_start_codon and t.contains_stop_codon:
coding_sequence_positions = tboundaries2positions(t)
if len(coding_sequence_positions) == len(t.coding_sequence):
#TODO need to check if the seq made from coding_sequence_positions is same as t.coding_seqeunce
dcoding = t2pmapper(t, coding_sequence_positions)
dcodingmutpos = dcoding.loc[(
dcoding['protein index'] == din.loc[
i, 'aminoacid: position']), :]
codon_positions = dcodingmutpos[
'coding sequence positions'].tolist()
if len(codon_positions) != 0:
dbed.loc[bedrowi, 'chromosome'] = t.contig
if cfg['test']:
print(din.loc[i, 'transcript: id'],
codon_positions)
if t.strand == '+':
dbed.loc[bedrowi,
'codon start'] = codon_positions[0]
dbed.loc[bedrowi, 'codon end'] = codon_positions[2]
elif t.strand == '-':
dbed.loc[bedrowi,
'codon start'] = codon_positions[2]
dbed.loc[bedrowi, 'codon end'] = codon_positions[0]
dbed.loc[bedrowi, 'start'] = dbed.loc[
bedrowi,
'codon start'] - 22 #FIXME put flank in the yml
dbed.loc[bedrowi, 'end'] = dbed.loc[
bedrowi,
'codon end'] + 21 #FIXME put flank in the yml
dbed.loc[bedrowi, 'reference residue'] = dcodingmutpos[
'protein sequence'].tolist()[0]
dbed.loc[bedrowi, 'reference codon'] = ''.join(
dcodingmutpos['coding sequence'].tolist())
dbed.loc[bedrowi, 'strand'] = t.strand
dbed.loc[bedrowi, 'id'] = '{}|{}|{}|{}|{}'.format(
din.loc[i, 'transcript: id'],
dbed.loc[bedrowi, 'chromosome'],
dbed.loc[bedrowi, 'strand'],
int(dbed.loc[bedrowi, 'start']),
int(dbed.loc[bedrowi, 'end']))
dbed.loc[bedrowi, 'gene: id'] = t.gene_id
dbed.loc[bedrowi, 'gene: name'] = t.gene.name
dbed.loc[bedrowi, 'protein: id'] = t.protein_id
dbed.loc[bedrowi, 'aminoacid: position'] = din.loc[
i, 'aminoacid: position']
# break
bedrowi += 1
else:
terrpositions.append(t.id)
else:
terrpositions.append(t.id)
else:
terrnoncoding.append(t.id)
else:
terrnotfound.append(din.loc[i, 'transcript: id'])
if cfg['test']:
logging.error('not found: {}'.format(
din.loc[i, 'transcript: id']))
if len(dbed) == 0:
from beditor.lib.global_vars import saveemptytable
logging.warning('no valid seqeunces found; saving an empty table.')
saveemptytable(cfg, f"{cfg['dsequencesp']}")
return None
dbed = dbed.loc[(dbed.apply(lambda x: x['end'] - x['start'] == 45,
axis=1)), :] #FIXME put flank in the yml
dbed.loc[:, 'start'] = dbed.loc[:, 'start'].astype(int)
dbed.loc[:, 'end'] = dbed.loc[:, 'end'].astype(int)
dbed = dbed.drop_duplicates(subset=bed_colns)
dbed.loc[:, bed_colns].to_csv(dbedp, sep='\t', header=False, index=False)
err2tids = {
'terrpositions': terrpositions,
'terrnotfound': terrnotfound,
'terrnoncoding': terrnoncoding,
}
if cfg['test']:
print(err2tids)
with open(dbedp + '.err.json', 'w') as outfile:
json.dump(err2tids, outfile)
bedp = f"{cfg['datad']}/dbedflank.bed"
fastap = f"{cfg['datad']}/dbedflank.fa"
cmd = f"{cfg['bedtools']} getfasta -s -name -fi {cfg['genomep']} -bed {bedp} -fo {fastap}"
runbashcmd(cmd)
dflankfa = fa2df(fastap, ids2cols=True)
dflankfa.loc[:, 'sequence'] = dflankfa.loc[:, 'sequence'].apply(
lambda x: x.upper())
dflankfa.loc[:,
'sequence: length'] = [len(s) for s in dflankfa['sequence']]
dflankfa.index = [idx.split('(')[0] for idx in dflankfa.index]
dflankfa.index.name = 'id'
dseq = set_index(dbed, 'id').join(set_index(dflankfa, 'id'), rsuffix='.1')
dseq2compatible = {
'aminoacid: position': 'aminoacid: position',
'gene: id': 'gene: id',
'gene: name': 'gene: name',
'protein: id': 'protein: id',
'transcript: id': 'seqid',
'transcript: sequence': 'sequence',
'aminoacid: wild-type': 'reference residue',
'codon: wild-type': 'reference codon',
'contig': 'contig',
'strand': 'strand',
'start': 'start',
'end': 'end',
'codon start': 'codon start',
'codon end': 'codon end',
}
if 'amino acid mutation' in dseq:
dseq2compatible['amino acid mutation'] = 'amino acid mutation'
dseq.to_csv(cfg['dseqtmpp'], sep='\t')
dseq = dseq[list(dseq2compatible.values())]
dseq.columns = list(dseq2compatible.keys())
# dseq.to_csv('data/dseq.csv')
logging.info(dseq.columns.tolist())
logging.info(din.columns.tolist())
dseq = pd.merge(dseq.reset_index(),
din,
on=['transcript: id', 'aminoacid: position'])
logging.info(dseq.columns.tolist())
set_index(dseq, 'id')
if 'reverse_mutations' in cfg:
if cfg['reverse_mutations']:
from beditor.lib.io_dfs import dfswapcols
dseq = dfswapcols(dseq,
['aminoacid: wild-type', 'amino acid mutation'])
dseq['codon: mutation'] = dseq['codon: wild-type'].copy()
dseq.to_csv(f"{cfg['dsequencesp']}", sep='\t')
del ensembl
def dguides2offtargets(cfg):
"""
All the processes in offtarget detection are here.
:param cfg: Configuration settings provided in .yml file
"""
from beditor.lib.global_vars import saveemptytable
cfg['datad'] = cfg[cfg['step']]
cfg['plotd'] = cfg['datad']
dofftargetsp = '{}/dofftargets.tsv'.format(cfg['datad'])
stepn = '04_offtargets'
logging.info(stepn)
dguidesp = f"{cfg[cfg['step']-1]}/dguides.tsv"
if not exists(dguidesp):
logging.warning(f"not found {dguidesp}")
return saveemptytable(cfg, dofftargetsp)
dguides = pd.read_csv(dguidesp, sep='\t')
if len(dguides) == 0:
logging.warning(f"dguides is empty.")
return saveemptytable(cfg, dofftargetsp)
cfg['datatmpd'] = f"{cfg['datad']}/tmp"
for dp in [cfg['datatmpd']]:
if not exists(dp):
makedirs(dp)
step2doutp = {
1: '01_guides_guidel*.fa',
2: '02_dalignbed.tsv',
3: '03_annotations.bed',
4: '04_dalignbedguides.tsv',
5: '05_dalignedfasta.tsv',
6: '06_dalignbedguidesseq.tsv',
7: '07_dalignbedstats.tsv',
8: '08_dannotsagg.tsv',
9: '09_dalignbedannot.tsv',
10: '10_daggbyguide.tsv',
}
cfg['dguidesp'] = dguidesp
cfg['alignmentbedp'] = f"{cfg['datatmpd']}/02_alignment.bed"
cfg['dalignbedp'] = f"{cfg['datatmpd']}/02_dalignbed.tsv"
cfg['dalignbedguidesp'] = f"{cfg['datatmpd']}/04_dalignbedguides.tsv"
cfg['dalignedfastap'] = f"{cfg['datatmpd']}/05_dalignedfasta.tsv"
cfg['dalignbedguidesseqp'] = f"{cfg['datatmpd']}/06_dalignbedguidesseq.tsv"
cfg['dalignbedstatsp'] = f"{cfg['datatmpd']}/07_dalignbedstats.tsv"
cfg['dannotsaggp'] = f"{cfg['datatmpd']}/08_dannotsagg.tsv"
cfg['dalignbedannotp'] = f"{cfg['datatmpd']}/09_dalignbedannot.tsv"
cfg['daggbyguidep'] = f"{cfg['datatmpd']}/10_daggbyguide.tsv"
#check which step to process
for step in range(2, 10 + 1, 1):
if not exists(f"{cfg['datatmpd']}/{step2doutp[step]}"):
if step == 2:
step = 'all'
break
logging.info(f'process from step:{step}')
cfg['dofftargetsp'] = '{}/dofftargets.tsv'.format(cfg['datad'])
if not exists(cfg['dofftargetsp']) or cfg['force']:
if step == 1 or step == 'all':
cfg = dguides2guidessam(cfg, dguides)
if step == 2 or step == 'all' or (not cfg is None):
cfg = guidessam2dalignbed(cfg)
if step == 3 or step == 'all' or (not cfg is None):
cfg = dalignbed2annotationsbed(cfg)
if step == 4 or step == 'all' or (not cfg is None):
cfg = dalignbed2dalignbedguides(cfg)
if step == 5 or step == 'all' or (not cfg is None):
cfg = alignmentbed2dalignedfasta(cfg)
if step == 6 or step == 'all' or (not cfg is None):
cfg = dalignbed2dalignbedguidesseq(cfg)
if step == 7 or step == 'all' or (not cfg is None):
cfg = dalignbedguidesseq2dalignbedstats(cfg)
if step == 8 or step == 'all' or (not cfg is None):
cfg = dannots2dalignbed2dannotsagg(cfg)
if step == 9 or step == 'all' or (not cfg is None):
cfg = dannotsagg2dannots2dalignbedannot(cfg)
if step == 10 or step == 'all' or (not cfg is None):
cfg = dalignbedannot2daggbyguide(cfg)
if cfg is None:
logging.warning(f"no alignment found")
cfg['step'] = 4
return saveemptytable(cfg, cfg['dofftargetsp'])
import gc
gc.collect()
def get_possible_mutagenesis(
cfg,
dcodontable,
dcodonusage,
BEs,
pos_muts,
host,
):
"""
Assesses possible mutagenesis strategies, given the set of Base editors and positions of mutations.
:param dcodontable: Codon table
:param dcodonusage: Codon usage table
:param BEs: Base editors (dict), see global_vars.py
:param pos_muts: positions of mutations
:param host: host organism
:returns: possible mutagenesis strategies as a pandas dataframe
"""
def write_dmutagenesis(cdni, posi, codon, codonmut, ntwt, ntmut, aa, aamut,
method):
"""
Write dmutagenesis table for each iteraction in get_possible_mutagenesis.
"""
dmutagenesis.loc[cdni, 'codon'] = codon
dmutagenesis.loc[cdni, 'position of mutation in codon'] = int(posi)
dmutagenesis.loc[cdni, 'codon mutation'] = codonmut
dmutagenesis.loc[cdni, 'nucleotide'] = ntwt
dmutagenesis.loc[cdni, 'nucleotide mutation'] = ntmut
dmutagenesis.loc[cdni, 'amino acid'] = aa
dmutagenesis.loc[cdni, 'amino acid mutation'] = aamut
dmutagenesis.loc[cdni, 'mutation on strand'] = method.split(' on ')[1]
dmutagenesis.loc[cdni,
'strand: mutation'] = method.split(' on ')[1].replace(
' strand', '')
dmutagenesis.loc[cdni, 'method'] = method.split(' on ')[0]
dmutagenesis.loc[cdni,
'codon mutation usage Fraction'] = dcodonusage.loc[
codonmut, 'Fraction']
dmutagenesis.loc[cdni,
'codon mutation usage Frequency'] = dcodonusage.loc[
codonmut, 'Frequency']
return dmutagenesis
def get_sm(dmutagenesis, BEs, positions, codon, muti, cdni):
"""
Fetches single nucleotide mutagenesis strategies.
"""
for method in BEs:
for posi in positions:
if BEs[method][0] == codon[posi]:
for ntmut in BEs[method][1]:
if posi == 0:
codonmut = '{}{}{}'.format(ntmut, codon[1],
codon[2])
elif posi == 1:
codonmut = '{}{}{}'.format(codon[0], ntmut,
codon[2])
elif posi == 2:
codonmut = '{}{}{}'.format(codon[0], codon[1],
ntmut)
aamut = str(
Seq.Seq(codonmut,
Alphabet.generic_dna).translate(table=1))
# if (aamut!='*') and (aamut!=aa): # nonsence and synonymous
if muti == 0:
cdni = cdni
else:
cdni = len(dmutagenesis) + 1
muti += 1
ntwt = BEs[method][0]
if '-' in method.split(' on ')[1]:
ntwt = str(
Seq.Seq(
ntwt,
Alphabet.generic_dna).reverse_complement())
ntmut = str(
Seq.Seq(
ntmut,
Alphabet.generic_dna).reverse_complement())
dmutagenesis_row = {
'cdni': cdni,
'posi': posi + 1,
'codon': codon,
'codonmut': codonmut,
'ntwt': ntwt,
'ntmut': ntmut,
'aa': aa,
'aamut': aamut,
'method': method
}
# print(dmutagenesis_row)
dmutagenesis = write_dmutagenesis(**dmutagenesis_row)
# else:
# logging.warning(f"BEs[{method}][0]!=codon[{posi}]")
return dmutagenesis, muti
#double nucleotide mutations
positions = {0: '@1st position', 1: '@2nd position', 2: '@3rd position'}
#double nucleotide mutations
positions_dm = [(i, j) for i in positions.keys() for j in positions.keys()
if i < j]
#double nucleotide mutations
positions_tm = [[0, 1, 2]]
dmutagenesis = dcodontable.copy()
# test=True
test = False
for cdni in dmutagenesis.index:
codon = dmutagenesis.loc[cdni, 'codon']
aa = dmutagenesis.loc[cdni, 'amino acid']
muti = 0
if test:
print(codon)
#single nucleuotide mutations
dmutagenesis, muti = get_sm(dmutagenesis, BEs, positions, codon, muti,
cdni)
if len(dmutagenesis) == 0:
from beditor.lib.global_vars import saveemptytable
logging.warning('no guides designed; saving an empty table.')
dmutagenesis = saveemptytable(cfg)
else:
# to_table(dmutagenesis,'test.tsv') #FIXME
# print(dmutagenesis.shape)
dmutagenesis = dmutagenesis.dropna() #FIXME
# print(dmutagenesis.shape)
dmutagenesis['nucleotide mutation: count'] = [
len(s) for s in dmutagenesis['nucleotide mutation']
]
dmutagenesis = dmutagenesis.sort_values('codon')
# Adding information of Allowed activity window
dmutagenesis = dmutagenesis.set_index('method').join(pos_muts)
dmutagenesis = dmutagenesis.reset_index()
from beditor.lib.io_seqs import reverse_complement_multintseq
from beditor.lib.global_vars import nt2complement
dmutagenesis['nucleotide: wild-type'] = dmutagenesis.apply(
lambda x: x['nucleotide']
if x['strand: mutation'] == '+' else reverse_complement_multintseq(
x['nucleotide'], nt2complement),
axis=1)
dmutagenesis['nucleotide: mutation'] = dmutagenesis.apply(
lambda x: x['nucleotide mutation']
if x['strand: mutation'] == '+' else reverse_complement_multintseq(
x['nucleotide mutation'], nt2complement),
axis=1)
return dmutagenesis
def get_possible_mutagenesis(
cfg,
dcodontable,
dcodonusage,
BEs,
pos_muts,
host,
):
"""
Assesses possible mutagenesis strategies, given the set of Base editors and positions of mutations.
:param dcodontable: Codon table
:param dcodonusage: Codon usage table
:param BEs: Base editors (dict), see global_vars.py
:param pos_muts: positions of mutations
:param host: host organism
:returns: possible mutagenesis strategies as a pandas dataframe
"""
def write_dmutagenesis(cdni, posi, codon, codonmut, ntwt, ntmut, aa, aamut,
method):
"""
Write dmutagenesis table for each iteraction in get_possible_mutagenesis.
"""
dmutagenesis.loc[cdni, 'codon'] = codon
dmutagenesis.loc[cdni, 'position of mutation in codon'] = int(posi)
dmutagenesis.loc[cdni, 'codon mutation'] = codonmut
dmutagenesis.loc[cdni, 'nucleotide'] = ntwt
dmutagenesis.loc[cdni, 'nucleotide mutation'] = ntmut
dmutagenesis.loc[cdni, 'amino acid'] = aa
dmutagenesis.loc[cdni, 'amino acid mutation'] = aamut
dmutagenesis.loc[cdni, 'mutation on strand'] = method.split(' on ')[1]
dmutagenesis.loc[cdni,
'strand: mutation'] = method.split(' on ')[1].replace(
' strand', '')
dmutagenesis.loc[cdni, 'method'] = method.split(' on ')[0]
dmutagenesis.loc[cdni,
'codon mutation usage Fraction'] = dcodonusage.loc[
codonmut, 'Fraction']
dmutagenesis.loc[cdni,
'codon mutation usage Frequency'] = dcodonusage.loc[
codonmut, 'Frequency']
return dmutagenesis
def get_sm(dmutagenesis, BEs, positions, codon, muti, cdni):
"""
Fetches single nucleotide mutagenesis strategies.
"""
for method in BEs:
for posi in positions:
if BEs[method][0] == codon[posi]:
for ntmut in BEs[method][1]:
if posi == 0:
codonmut = '{}{}{}'.format(ntmut, codon[1],
codon[2])
elif posi == 1:
codonmut = '{}{}{}'.format(codon[0], ntmut,
codon[2])
elif posi == 2:
codonmut = '{}{}{}'.format(codon[0], codon[1],
ntmut)
aamut = str(
Seq.Seq(codonmut,
Alphabet.generic_dna).translate(table=1))
# if (aamut!='*') and (aamut!=aa): # nonsence and synonymous
if muti == 0:
cdni = cdni
else:
cdni = len(dmutagenesis) + 1
muti += 1
ntwt = BEs[method][0]
if '-' in method.split(' on ')[1]:
ntwt = str(
Seq.Seq(
ntwt,
Alphabet.generic_dna).reverse_complement())
ntmut = str(
Seq.Seq(
ntmut,
Alphabet.generic_dna).reverse_complement())
dmutagenesis = write_dmutagenesis(
**{
'cdni': cdni,
'posi': posi + 1,
'codon': codon,
'codonmut': codonmut,
'ntwt': ntwt,
'ntmut': ntmut,
'aa': aa,
'aamut': aamut,
'method': method
})
return dmutagenesis, muti
def get_dm(dmutagenesis, BEs, positions_dm, codon, muti, cdni):
"""
Fetches double nucleotide mutagenesis strategies.
"""
for method in BEs:
for posi1, posi2 in positions_dm:
if (BEs[method][0] == codon[posi1]) and (BEs[method][0]
== codon[posi2]):
for ntmut1, ntmut2 in itertools.product(''.join(
BEs[method][1]),
repeat=2):
if (posi1 == 0) and (posi2 == 1):
codonmut = '{}{}{}'.format(ntmut1, ntmut2,
codon[2])
elif (posi1 == 1) and (posi2 == 2):
codonmut = '{}{}{}'.format(codon[0], ntmut1,
ntmut2)
elif (posi1 == 0) and (posi2 == 2):
codonmut = '{}{}{}'.format(ntmut1, codon[1],
ntmut2)
aamut = str(
Seq.Seq(codonmut,
Alphabet.generic_dna).translate(table=1))
# if (aamut!='*') and (aamut!=aa): # nonsence and synonymous
if muti == 0:
cdni = cdni
else:
cdni = len(dmutagenesis) + 1
muti += 1
ntwt = '{}{}'.format(BEs[method][0], BEs[method][0])
ntmut = '{}{}'.format(ntmut1, ntmut2)
if '-' in method.split(' on ')[1]:
ntwt = str(
Seq.Seq(
ntwt,
Alphabet.generic_dna).reverse_complement())
ntmut = str(
Seq.Seq(
ntmut,
Alphabet.generic_dna).reverse_complement())
dmutagenesis = write_dmutagenesis(
**{
'cdni': cdni,
'posi': '{}{}'.format(posi1, posi2),
'codon': codon,
'codonmut': codonmut,
'ntwt': ntwt,
'ntmut': ntmut,
'aa': aa,
'aamut': aamut,
'method': method
})
return dmutagenesis, muti
def get_tm(dmutagenesis, BEs, positions_tm, codon, muti, cdni):
"""
Fetches triple nucleotide mutagenesis strategies.
"""
for method in BEs:
for posi1, posi2, posi3 in positions_tm:
if (BEs[method][0] == codon[posi1]) and (
BEs[method][0] == codon[posi2]) and (BEs[method][0]
== codon[posi3]):
for ntmut1, ntmut2, ntmut3 in itertools.product(''.join(
BEs[method][1]),
repeat=3):
codonmut = '{}{}{}'.format(ntmut1, ntmut2, ntmut3)
aamut = str(
Seq.Seq(codonmut,
Alphabet.generic_dna).translate(table=1))
# if (aamut!='*') and (aamut!=aa): # nonsence and synonymous
if muti == 0:
cdni = cdni
else:
cdni = len(dmutagenesis) + 1
muti += 1
ntwt = '{}{}{}'.format(BEs[method][0], BEs[method][0],
BEs[method][0])
ntmut = '{}{}{}'.format(ntmut1, ntmut2, ntmut3)
if '-' in method.split(' on ')[1]:
ntwt = str(
Seq.Seq(
ntwt,
Alphabet.generic_dna).reverse_complement())
ntmut = str(
Seq.Seq(
ntmut,
Alphabet.generic_dna).reverse_complement())
dmutagenesis = write_dmutagenesis(
**{
'cdni': cdni,
'posi': '123',
'codon': codon,
'codonmut': codonmut,
'ntwt': ntwt,
'ntmut': ntmut,
'aa': aa,
'aamut': aamut,
'method': method
})
return dmutagenesis, muti
def get_dm_combo(dmutagenesis, BEs, positions_dm, codon, muti, cdni,
method):
"""
Fetches double nucleotide mutagenesis strategies utilising 2 different base editors simultaneously.
"""
methods = [
m for m in itertools.product(BEs.keys(), repeat=2) if ((
m[0].split('on')[1] == m[1].split('on')[1])) and (m[0] != m[1])
]
for method1, method2 in methods:
for posi1, posi2 in positions_dm:
if (BEs[method1][0] == codon[posi1]) and (BEs[method2][0]
== codon[posi2]):
ntmuts = [(n1, n2) for n1 in ''.join(BEs[method1][1])
for n2 in ''.join(BEs[method2][1])]
for ntmut1, ntmut2 in ntmuts:
if (posi1 == 0) and (posi2 == 1):
codonmut = '{}{}{}'.format(ntmut1, ntmut2,
codon[2])
elif (posi1 == 1) and (posi2 == 2):
codonmut = '{}{}{}'.format(codon[0], ntmut1,
ntmut2)
elif (posi1 == 0) and (posi2 == 2):
codonmut = '{}{}{}'.format(ntmut1, codon[1],
ntmut2)
aamut = str(
Seq.Seq(codonmut,
Alphabet.generic_dna).translate(table=1))
# if (aamut!='*') and (aamut!=aa): # nonsence and synonymous
if muti == 0:
cdni = cdni
else:
cdni = len(dmutagenesis) + 1
muti += 1
ntwt = '{}{}'.format(BEs[method1][0], BEs[method2][0])
ntmut = '{}{}'.format(ntmut1, ntmut2)
if '-' in method1.split(' on ')[1]:
ntwt = str(
Seq.Seq(
ntwt,
Alphabet.generic_dna).reverse_complement())
ntmut = str(
Seq.Seq(
ntmut,
Alphabet.generic_dna).reverse_complement())
dmutagenesis = write_dmutagenesis(
**{
'cdni': cdni,
'posi': '{}{}'.format(posi1, posi2),
'codon': codon,
'codonmut': codonmut,
'ntwt': ntwt,
'ntmut': ntmut,
'aa': aa,
'aamut': aamut,
'method': method + ' on ' +
method1.split('on')[1]
})
return dmutagenesis, muti
def get_tm_combo(dmutagenesis, BEs, positions_tm, codon, muti, cdni,
method):
"""
Fetches triple nucleotide mutagenesis strategies utilising 2 different base editors simultaneously.
"""
methods = [
m for m in itertools.product(BEs.keys(), repeat=3)
if ((m[0].split('on')[1] == m[1].split('on')[1] == m[2].split('on')
[1])) and not (m[0] == m[1] == m[2])
]
for method1, method2, method3 in methods:
for posi1, posi2, posi3 in positions_tm:
if (BEs[method1][0] == codon[posi1]) and (
BEs[method2][0] == codon[posi2]) and (BEs[method3][0]
== codon[posi3]):
ntmuts = [(n1, n2, n3) for n1 in ''.join(BEs[method1][1])
for n2 in ''.join(BEs[method2][1])
for n3 in ''.join(BEs[method3][1])]
for ntmut1, ntmut2, ntmut3 in ntmuts:
codonmut = '{}{}{}'.format(ntmut1, ntmut2, ntmut3)
aamut = str(
Seq.Seq(codonmut,
Alphabet.generic_dna).translate(table=1))
# if (aamut!='*') and (aamut!=aa): # nonsence and synonymous
if muti == 0:
cdni = cdni
else:
cdni = len(dmutagenesis) + 1
muti += 1
ntwt = '{}{}{}'.format(BEs[method1][0],
BEs[method2][0],
BEs[method3][0])
ntmut = '{}{}{}'.format(ntmut1, ntmut2, ntmut3)
if '-' in method1.split(' on ')[1]:
ntwt = str(
Seq.Seq(
ntwt,
Alphabet.generic_dna).reverse_complement())
ntmut = str(
Seq.Seq(
ntmut,
Alphabet.generic_dna).reverse_complement())
dmutagenesis = write_dmutagenesis(
**{
'cdni': cdni,
'posi': '123',
'codon': codon,
'codonmut': codonmut,
'ntwt': ntwt,
'ntmut': ntmut,
'aa': aa,
'aamut': aamut,
'method': method + ' on ' +
method1.split('on')[1]
})
return dmutagenesis, muti
#double nucleotide mutations
positions = {0: '@1st position', 1: '@2nd position', 2: '@3rd position'}
#double nucleotide mutations
positions_dm = [(i, j) for i in positions.keys() for j in positions.keys()
if i < j]
#double nucleotide mutations
positions_tm = [[0, 1, 2]]
dmutagenesis = dcodontable.copy()
# test=True
test = False
for cdni in dmutagenesis.index:
codon = dmutagenesis.loc[cdni, 'codon']
aa = dmutagenesis.loc[cdni, 'amino acid']
muti = 0
if test:
print(codon)
#single nucleuotide mutations
dmutagenesis, muti = get_sm(dmutagenesis, BEs, positions, codon, muti,
cdni)
#double nucleotide mutations
dmutagenesis, muti = get_dm(dmutagenesis, BEs, positions_dm, codon,
muti, cdni)
#triple nucleotide mutations
dmutagenesis, muti = get_tm(dmutagenesis, BEs, positions_tm, codon,
muti, cdni)
# #double nucleotide mutations combinations
# dmutagenesis,muti=get_dm_combo(dmutagenesis,BEs,positions_dm,codon,muti,cdni,method='undefined')
# #triple nucleotide mutations combinations
# dmutagenesis,muti=get_tm_combo(dmutagenesis,BEs,positions_tm,codon,muti,cdni,method='undefined')
if len(dmutagenesis) == 0:
from beditor.lib.global_vars import saveemptytable
logging.warning('no guides designed; saving an empty table.')
dmutagenesis = saveemptytable(cfg)
else:
dmutagenesis['nucleotide mutation: count'] = [
len(s) for s in dmutagenesis['nucleotide mutation']
]
dmutagenesis = dmutagenesis.sort_values('codon')
# Adding information of Allowed activity window
dmutagenesis = dmutagenesis.set_index('method').join(pos_muts)
dmutagenesis = dmutagenesis.reset_index()
from beditor.lib.io_seqs import reverse_complement_multintseq
from beditor.lib.global_vars import nt2complement
dmutagenesis['nucleotide: wild-type'] = dmutagenesis.apply(
lambda x: x['nucleotide']
if x['strand: mutation'] == '+' else reverse_complement_multintseq(
x['nucleotide'], nt2complement),
axis=1)
dmutagenesis['nucleotide: mutation'] = dmutagenesis.apply(
lambda x: x['nucleotide mutation']
if x['strand: mutation'] == '+' else reverse_complement_multintseq(
x['nucleotide mutation'], nt2complement),
axis=1)
return dmutagenesis
def dseq2dguides(cfg):
"""
Wrapper around make guides function.
:param cfg: configuration dict.
"""
cfg['datad'] = cfg[cfg['step']]
cfg['plotd'] = cfg['datad']
dguideslinp = f"{cfg['datad']}/dguides.tsv"
dguides_nofltp = f"{cfg['datad']}/dguides_noflt.tsv"
dmutagenesisp = f"{cfg['datad']}/dmutagenesis.tsv"
if not exists(dguideslinp) or cfg['force']:
dmutagenesis = pd.read_csv(f"{cfg[cfg['step']-1]}/dmutagenesis.tsv",
sep='\t',
keep_default_na=False)
if cfg['mutation_format'] == 'nucleotide':
dsequences = pd.read_csv(
f"{cfg[cfg['step']-2]}/dsequences.tsv",
sep='\t',
keep_default_na=False
) #FIXME if numbering of steps is changed, this is gonna blow
dsequences, dmutagenesis = dinnucleotide2dsequencesproper(
dsequences, dmutagenesis)
elif cfg['mutation_format'] == 'aminoacid':
dsequences = pd.read_csv(
f"{cfg[cfg['step']-2]}/dsequences.tsv",
sep='\t',
keep_default_na=False
) #FIXME if numbering of steps is changed, this is gonna blow
if 'reverse_mutations' in cfg:
if cfg['reverse_mutations']:
# from beditor.lib.global_vars import stepi2cols
cols_dsequences = dsequences.columns.tolist()
dsequences = pd.merge(dsequences,
dmutagenesis,
how='inner',
left_on=[
'aminoacid: wild-type',
'codon: mutation',
'amino acid mutation'
],
right_on=[
'amino acid', 'codon mutation',
'amino acid mutation'
],
suffixes=['', ': dmutagenesis'])
dsequences['codon: wild-type'] = dsequences['codon']
dsequences = dsequences.loc[:, cols_dsequences]
dsequences.to_csv(f"{cfg[cfg['step']]}/dsequences.tsv", sep='\t')
if not (len(dsequences) == 0 or len(dmutagenesis) == 0):
dmutagenesis['strand'] = dmutagenesis.apply(
lambda x: x['mutation on strand'].replace(' strand', ''),
axis=1)
dmutagenesis.to_csv(dmutagenesisp, sep='\t')
dguideslin, dguides_noflt, err2idxs, dguides_neg_control, dguides_pos_control = make_guides(
cfg,
dsequences,
dmutagenesis,
test=cfg['test'],
# dbug=True,
)
if not dguides_noflt is None:
dguides_noflt.to_csv(dguides_nofltp, sep='\t')
if not ((dguideslin is None) and (err2idxs is None)):
dguideslin.to_csv(dguideslinp, sep='\t')
if cfg['test']:
logging.info(err2idxs)
with open(dguideslinp + '.err.json', 'w') as f:
json.dump(err2idxs, f)
if cfg['make_control_pos'] and not dguides_pos_control is None:
to_table(dguides_pos_control,
f"{dguideslinp}.pos_control.tsv")
if cfg['make_control_neg'] and not dguides_neg_control is None:
to_table(dguides_neg_control,
f"{dguideslinp}.neg_control.tsv")
else:
from beditor.lib.global_vars import saveemptytable
logging.warning('no guides designed; saving an empty table.')
saveemptytable(cfg, dguideslinp)
else:
from beditor.lib.global_vars import saveemptytable
logging.warning('no guides designed; saving an empty table.')
saveemptytable(cfg, dguideslinp)
import gc
gc.collect()
