def get_rsid_list():
snp_dir = get_path("vertex/SNP")
snp_fn = "osu19_SNP.bed"
rsid_list = pd.read_csv(os.path.join(snp_dir, snp_fn),
sep="\t",
header=None,
usecols=[3],
names=["name"],
squeeze=True).tolist()
return rsid_list
def read_coexpedia_dfms(_dir):
for fn in os.listdir(_dir):
path = os.path.join(_dir, fn)
dfm = pd.read_csv(path,
sep="\t",
header=None,
names=["Gene_A", "Gene_B", "LLS"])
geo_accession_id = fn.split(".txt")[0]
dfm = dfm.assign(GEO_ACCESSION=geo_accession_id)
yield dfm
if __name__ == "__main__":
res_dir = get_path("resource/Coexpedia")
dfms = read_coexpedia_dfms(os.path.join(res_dir, "Hsa"))
merged_dfm = pd.concat(dfms, axis=0)
merged_dfm.to_csv(os.path.join(res_dir, "p1_coexpedia_merged.tsv"),
sep="\t",
index=False)
reduced_dfm = merged_dfm.loc[:, ["Gene_A", "Gene_B"]]
reduced_dfm = remove_duplicated_undirected_edges(reduced_dfm, sort=True)
edge_dir = get_path("edge/gene-gene")
reduced_dfm.to_csv(os.path.join(edge_dir, "Coexpedia.edgelist"),
sep="\t",
index=False,
header=False)
return df_a_minus_b
def output_snp_gene_edgelist(snp_gene_map):
snp_gene_el = snp_gene_map.loc[:, ["rs_id", "gene_ensembl_id"]].drop_duplicates()
snp_gene_el = map_Ensembl_IDs_to_Entrez(snp_gene_el, ensembl_colname="gene_ensembl_id", new_colname="gene_id", keep_unmapped=True)
snp_gene_el.sort_values(by="gene_id", inplace=True)
return snp_gene_el
def output_snp_snp_edgelist(snp_snp_map):
return snp_snp_map.loc[:, ["rs_id_A", "rs_id_B"]].drop_duplicates()
if __name__ == "__main__":
_4DGenome_dir = get_path("resource/4DGenome")
snp_in_interactorA = get_interacted_SNPs(os.path.join(_4DGenome_dir, "InteractorA_SNP_intxn.bed"))
snp_in_interactorB = get_interacted_SNPs(os.path.join(_4DGenome_dir, "InteractorB_SNP_intxn.bed"))
gene_tss_in_interactorA = get_genes_with_interacted_TSSs(os.path.join(_4DGenome_dir, "InteractorA_Ensembl_TSS_intxn.bed"))
gene_tss_in_interactorB = get_genes_with_interacted_TSSs(os.path.join(_4DGenome_dir, "InteractorB_Ensembl_TSS_intxn.bed"))
gene_prm_in_interactorA = get_genes_with_interacted_promoters(os.path.join(_4DGenome_dir, "InteractorA_Ensembl_promoter_intxn.bed"))
gene_prm_in_interactorB = get_genes_with_interacted_promoters(os.path.join(_4DGenome_dir, "InteractorB_Ensembl_promoter_intxn.bed"))
# interacted (SNP, gene) pairs (via TSSs/promoters)
snp_tss_map = _cross_inner_join(snp_in_interactorA, snp_in_interactorB, gene_tss_in_interactorA, gene_tss_in_interactorB)
snp_prm_map = _cross_inner_join(snp_in_interactorA, snp_in_interactorB, gene_prm_in_interactorA, gene_prm_in_interactorB)
# remove redundant (SNP, gene) pairs that are both associated by TSSs and promoters
eqtl_fn_list = list_files(eqtl_dir, eqtl_suffix)
for eqtl_fn in eqtl_fn_list:
eqtl_path = os.path.join(eqtl_dir, eqtl_fn)
eqtl_df = read_eqtl_df(eqtl_path)
snp_egene_map = snp_df.merge(eqtl_df, on="rs_id", how="inner")
eqtl_source = eqtl_fn.split(eqtl_suffix)[0]
snp_egene_map = snp_egene_map.assign(eqtl_source=eqtl_source)
yield snp_egene_map
if __name__ == "__main__":
snp_dir = get_path("vertex/SNP")
snp_fn = "osu19_SNP.bed"
snp_df = read_snp_df(os.path.join(snp_dir, snp_fn))
res_dir = get_path("resource/GTEx")
eqtl_dir = os.path.join(res_dir, "GTEx_Analysis_v7_eQTL")
eqtl_suffix = ".v7.egenes.txt"
snp_egene_map = pd.concat(gen_snp_egene_map(snp_df, eqtl_dir, eqtl_suffix),
ignore_index=True)
# Get rid of version numbers
snp_egene_map.loc[:,
"gene_ensembl_id"] = snp_egene_map.loc[:,
"gene_ensembl_id"].apply(
lambda x: x.
split(".")[0])
import os
import pandas as pd
from util_path import get_path
_gene_dir = get_path("vertex/gene")
_ensembl_entrez_map = pd.read_csv(os.path.join(_gene_dir,
"Ensembl_x_Entrez.tsv"),
sep="\t",
dtype={
"Ensembl_Gene_ID": str,
"Entrez_Gene_ID": str
})
# rename columns in order to lower the possibility of naming conflicts
_ensembl_entrez_map.rename(columns={
"Ensembl_Gene_ID": "_Ensembl_Gene_ID",
"Entrez_Gene_ID": "_Entrez_Gene_ID"
},
inplace=True)
def get_mapped_Ensembl_IDs():
return set(_ensembl_entrez_map._Ensembl_Gene_ID)
def map_Ensembl_IDs_to_Entrez(df,
ensembl_colname,
new_colname,
keep_unmapped=False):
"""
Please make sure that no column in `df` is named "_Ensembl_Gene_ID" or "_Entrez_Gene_ID"
"""
def get_rsid_list():
snp_dir = get_path("vertex/SNP")
snp_fn = "osu19_SNP.bed"
rsid_list = pd.read_csv(os.path.join(snp_dir, snp_fn),
sep="\t",
header=None,
usecols=[3],
names=["name"],
squeeze=True).tolist()
return rsid_list
if __name__ == "__main__":
res_dir = get_path("resource/EncodeTFBS")
rsid_list = get_rsid_list()
snp_tfbs_matrix = get_snp_tfbs_matrix(rsid_list, config_key="local_hg19")
snp_tfbs_matrix = update_tfbs_symbol(snp_tfbs_matrix)
snp_tfbs_matrix.to_csv(os.path.join(res_dir, "p1_SNP_x_TFBS_matrix.tsv"),
sep="\t",
header=True,
index=True)
snp_tfbs_map = convert_to_map(snp_tfbs_matrix,
map_colnames=["name", "symbol"])
snp_tfbs_map.to_csv(os.path.join(res_dir, "p2_SNP_x_TFBS_map.tsv"),
sep="\t",
header=True,
index=False)
import os
import pandas as pd
from util_path import get_path
from util_edge import remove_duplicated_undirected_edges
if __name__ == "__main__":
input_dir = get_path("resource/BioGRID")
edgelist_df = pd.read_csv(
os.path.join(input_dir, "BIOGRID-Human.tab2.tsv"),
sep="\t",
usecols=["Entrez Gene Interactor A", "Entrez Gene Interactor B"])
edgelist_df = remove_duplicated_undirected_edges(edgelist_df, sort=True)
output_dir = get_path("edge/gene-gene")
edgelist_df.to_csv(os.path.join(output_dir, "BioGRID.edgelist"),
sep="\t",
index=False,
header=False)
import os
import pandas as pd
from util_path import get_path
_dei_dir = get_path("resource/Util_DEI")
_dei_set = set(
pd.read_csv(os.path.join(_dei_dir, "discontinued_entrez_id.tsv"),
sep="\t",
usecols=["Discontinued_GeneID"],
dtype={
"Discontinued_GeneID": int
}).Discontinued_GeneID)
def filter_dei(df, id_col):
global _dei_set
found_dei = set(df[id_col]).intersection(_dei_set)
if found_dei:
print("[util_dei] Found Disctinued Entrez ID: {}".format(found_dei))
flag_dei = df[id_col].isin(found_dei)
return df.loc[~flag_dei, :]
import os
import pandas as pd
from util_path import get_path
from util_edge import remove_duplicated_undirected_edges
if __name__ == "__main__":
input_dir = get_path("resource/HumanNet")
edgelist_df = pd.read_csv(os.path.join(input_dir, "HumanNet-XN.tsv"),
sep="\t",
comment="#",
names=["Gene_A", "Gene_B", "LLS"],
usecols=["Gene_A", "Gene_B"])
edgelist_df = remove_duplicated_undirected_edges(edgelist_df, sort=True)
output_dir = get_path("edge/gene-gene")
edgelist_df.to_csv(os.path.join(output_dir, "HumanNet.edgelist"),
sep="\t",
index=False,
header=False)
def convert_to_df(closest_obj):
return pd.read_csv(StringIO(str(closest_obj)),
sep="\t",
header=None,
names=[
"snpChrom", "snpChromStart", "snpChromEnd",
"snpName", "geneChrom", "geneChromStart",
"geneChromEnd", "geneName", "geneID", "geneStrand",
"distance"
])
if __name__ == "__main__":
snp_dir = get_path("vertex/SNP")
snp_fn = "osu19_SNP.bed"
snp_start_bed = make_snp_start_BED(os.path.join(snp_dir, snp_fn))
gene_dir = get_path("vertex/gene")
ensembl_TSS_fn = "Ensembl_TSS.bed"
ensembl_TSS_start_bed = make_TSS_start_BED(
os.path.join(gene_dir, ensembl_TSS_fn))
entrez_TSS_fn = "Entrez_TSS.bed"
entrez_TSS_start_bed = make_TSS_start_BED(
os.path.join(gene_dir, entrez_TSS_fn))
snp_start_bt = convert_to_bedtool_obj(snp_start_bed)
ensembl_TSS_start_bt = convert_to_bedtool_obj(ensembl_TSS_start_bed)
entrez_TSS_start_bt = convert_to_bedtool_obj(entrez_TSS_start_bed)
import os
import pandas as pd
import mygene
from util_path import get_path
from util_dei import filter_dei
res_dir = get_path("resource/Entrez")
gene_dir = get_path("vertex/gene")
mg = mygene.MyGeneInfo()
def read_gene2ensembl():
global res_dir
g2e_df = pd.read_csv(os.path.join(res_dir, "gene2ensembl_9606.tsv"), sep="\t", header=None,
names=["Tax_ID", "Entrez_Gene_ID", "Ensembl_Gene_ID"])
unique_tax_ids = g2e_df.Tax_ID.unique()
assert(len(unique_tax_ids) == 1)
assert(unique_tax_ids[0] == 9606)
g2e_df = g2e_df.drop("Tax_ID", axis=1).drop_duplicates().reindex()
return g2e_df
def read_biomart(GRCh):
global res_dir
if GRCh == "37":
filename = "GRCh37_p13_mart_export.txt"
elif GRCh == "38":
