def get_pkis_pki(dataframe=True):
pkis = '../../kinome_assay/other_published/pkis_pki.tsv'
gene2uniprot, uniprot2genes, mutgene2mutuniprot, mutgene2filename = get_protein_idmap(
)
chembl2ikey = load_chembl2ikey()
data = []
with open(pkis, 'r') as f:
for line in f:
line = line.strip().split('\t')
chembl = line[0]
gene = line[1]
pki = float(line[2])
try:
ikey = chembl2ikey[chembl]
except:
#there may be outdated ChEMBL molecules removed from ChEMBL DB
continue
try:
uni = gene2uniprot[gene]
except:
#target genes not properly mapped to any protein
#with open('extra_genes_pkis.txt','a') as out:
# out.write("{}\n".format(gene))
continue
tup = (ikey, uni, 'pKi', '=', pki)
data.append(tup)
if dataframe:
data = pandas_df_continuous(data)
return data
def get_plos_pki(dataframe=True):
pkis = '../../kinome_assay/other_published/plos_pki.tsv'
gene2uniprot, uniprot2genes, mutgene2mutuniprot, mutgene2filename = get_protein_idmap(
)
data = []
with open(pkis, 'r') as f:
for line in f:
line = line.strip().split('\t')
ikey = line[0]
gene = line[1]
modobj = re.match(r'(.*)-(.*ted)$', gene, re.M | re.I)
if modobj:
gene = modobj.group(1)
mod = modobj.group(2)
else:
mod = None
pki = float(line[2])
try:
uni = gene2uniprot[gene]
except:
#target genes not properly mapped to any protein
#with open('extra_genes_plos.txt','a') as out:
# out.write("{}\n".format(gene))
continue
if mod is not None:
uni = uni + '-' + mod
tup = (ikey, uni, 'pKi', '=', pki)
data.append(tup)
if dataframe:
data = pandas_df_continuous(data)
return data
def get_jcim_pki(dataframe=True):
jcim = '../../kinome_assay/other_published/JCIM_activity_pKi.tsv'
gene2uniprot, uniprot2genes, mutgene2mutuniprot, mutgene2filename = get_protein_idmap(
)
chemname2ikey = load_chemname2ikey()
data = []
with open(jcim, 'r') as f:
next(f)
for line in f:
line_ori = line
line = line.strip().split('\t')
org = line[0]
gene = str(line[1]).replace(
' ', '') #genes in JCIM may contain white spaces
comp = str(line[2]) #compound name
pki = float(line[3]) #pKi in Ki (M)
try:
uni = gene2uniprot[gene]
except:
# with open('./extra_genes_jcim.txt','a') as out:
# out.write("{}\n".format(gene))
continue
try:
ikey = chemname2ikey[comp]
except:
with open('./extra_chemicals_jcim.txt', 'a') as out:
out.write("{}\n".format(comp))
continue
tup = (ikey, uni, 'pKi', '=', pki)
data.append(tup)
if dataframe:
data = pandas_df_continuous(data)
return data
def get_bindingdb_by_assay_type(assay_type='pKd',dataframe=True):
bindingdb_path='../../BindingDB/'
pic50=bindingdb_path+'BindingDB_pIC50.tsv'
pkd=bindingdb_path+'BindingDB_pKd.tsv'
pki=bindingdb_path+'BindingDB_pKi.tsv'
if (assay_type=='pIC50') or (assay_type=='pic50'):
infile=pic50
atype='pIC50'
elif (assay_type=='pKd') or (assay_type=='pkd'):
infile=pkd
atype='pKd'
elif (assay_type=='pKi') or (assay_type=='pki'):
infile=pki
atype='pKi'
else:
print("Error in parsing BindingDB data. Choose a proper assay type (pIC50, pKd, or pKi)")
sys.exit()
data=[]
with open(infile,'r') as f:
for line in f:
line=line.strip().split('\t')
ikey=str(line[0])
uni=str(line[1])
rel=line[2]
val=float(line[3])
tup=(ikey,uni,atype,rel,val)
data.append(tup)
if dataframe:
data=pandas_df_continuous(data)
return data
def get_chembl_cyp450_by_assay_type(assay_type='pKd',dataframe=True):
fpath='../../CYP450/ChEMBL23/'
pic50=fpath+'CYP450_pIC50.tsv'
pkd=fpath+'CYP450_pKd.tsv'
pki=fpath+'CYP450_pKi.tsv'
if (assay_type=='pIC50') or (assay_type=='pic50'):
infile=pic50
atype='pIC50'
elif (assay_type=='pKd') or (assay_type=='pkd'):
infile=pkd
atype='pKd'
elif (assay_type=='pKi') or (assay_type=='pki'):
infile=pki
atype='pKi'
else:
print("Error in parsing ChEMBL CYP450 data. Choose a proper assay type (pIC50, pKd, or pKi)")
sys.exit()
data=[]
with open(infile,'r') as f:
next(f)
for line in f:
line=line.strip().split('\t')
ikey=str(line[0])
uni=str(line[1])
rel=line[2]
val=float(line[3])
tup=(ikey,uni,atype,rel,val)
data.append(tup)
if dataframe:
data=pandas_df_continuous(data)
return data
def get_kinomescan(assay_type='pKd', dataframe=True):
#if dataframe=True, a pandas dataframe is returned
#dataframe has integer indice from 0,
# column names are ['InChIKey','UniProt','Activity_type','Relation','Activity_value']
gene2uniprot, uniprot2genes, mutgene2mutuniprot, mutgene2filename = get_protein_idmap(
)
chemname2ikey = load_chemname2ikey()
kinomepath = '../../kinome_assay/LINCS/'
if assay_type == 'pKd' or assay_type == 'pkd':
assay_type = 'pKd'
activity = kinomepath + 'LINCS_kinomescan_kd_nM.tsv' #assay with numeric activity value in Kd
null_activity = kinomepath + 'LINCS_kinomescan_kd_inactive_null.tsv' # assay inactive, without numeric value reported
elif assay_type == 'pi' or assay_type == 'pI':
assay_type = 'pPI' #Percent Inhibition Standardized = compound_concentration_nM*(100-%activity)/%activity
activity = kinomepath + 'LINCS_kinomescan_pi_nM.tsv' #assay with numeric activity value in PI
null_activity = kinomepath + 'LINCS_kinomescan_pi_inactive_null.tsv' # assay inactive, without numeric value reported
else:
print("Choose activity. pKd or pI.")
sys.exit()
data = []
with open(activity, 'r') as f:
for l in f:
l = l.strip().split('\t')
drug = l[0]
gene = l[1]
modobj = re.match(r'(.*)-(.*ted)$', gene, re.M | re.I)
if modobj:
gene = modobj.group(1)
mod = modobj.group(2)
else:
mod = None
val = float(l[2])
if val <= 0: #nonsense data, Kd must be positive
continue
val = -np.log10(val) + 9.0 #all activities are in nM
if np.isinf(val) or np.isnan(val):
continue #skip inf or NaN
ikey = chemname2ikey[drug]
try:
uni = gene2uniprot[gene]
except:
# with open('./extra_genes_in_kinomescan.txt','a') as out: #collect unmapped genes
# out.write("{}\n".format(gene))
continue
if mod is not None:
uni = uni + '-' + mod
tup = (ikey, uni, assay_type, '=', val)
data.append(tup)
if dataframe:
data = pandas_df_continuous(data)
return data
def get_gpcrdb(assay_type='pKd', dataframe=True):
#if dataframe=True, a pandas dataframe is returned
#dataframe has integer indice from 0,
# column names are ['InChIKey','UniProt','Activity_type','Relation','Activity_value']
fpath = '../../GPCRdb/'
gpcrfile = fpath + 'GPCR_assays.csv.gz'
if Path(gpcrfile).is_file(): #decompress if compressed
gpcrfile = decompress_gzip(gpcrfile)
else:
gpcrfile = fpath + 'GPCR_assays.csv' #already decompressed
chembl2uniprot = load_chembl2uniprot()
chembl2ikey = load_chembl2ikey()
data = []
with open(gpcrfile, 'r') as f:
#activity types : ['AC50', 'Potency', 'IC50', 'EC50', 'Kd', 'Ki']
#activity units : ['nM']
next(f)
for l in csv.reader(f,
quotechar='"',
delimiter=',',
quoting=csv.QUOTE_ALL,
skipinitialspace=True):
smi = l[8]
c_chembl = l[14] #chembl ID for chemical molecule
rel = l[25]
atype = l[26]
unit = l[27]
val = l[28]
atype = 'p' + atype #e.g. IC50 -> pIC50
if atype.lower() != assay_type.lower():
continue
p_chembl = l[29]
try:
ikey = chembl2ikey[c_chembl]
except:
continue
try:
uni = chembl2uniprot[p_chembl]
except:
continue
try:
val = np.float(val)
except:
continue
val = -np.log10(val) + 9.0 #all activities are in nM
if np.isinf(val) or np.isnan(val):
continue #skip inf or NaN
if rel != '=': #need to flip sign for -log conversion unless '='
lt = re.search(r'<', rel)
gt = re.search(r'>', rel)
if lt:
if gt: #relation shouldn't contain both > and <
continue
rel = rel.replace('<', '>')
elif gt:
rel = rel.replace('>', '<')
else:
continue
tup = (ikey, uni, atype, rel, val)
data.append(tup)
if dataframe:
data = pandas_df_continuous(data)
return data