h5ad_ID2symbol

import anndata as ad
import pandas as pd
import numpy as np
from gseapy import Biomart
adata = ad.read_h5ad("kidney_tumors.h5ad")

adata.var.head(2)

	vst.mean	vst.variance	vst.variance.expected	vst.variance.standardized	vst.variable	feature_is_filtered	feature_name	feature_reference	feature_biotype
ENSG00000243485	0.000216	0.000216	0.000216	0.999738	0	0	15855	0	0
ENSG00000237613	0.000000	0.000000	0.000000	0.000000	0	0	8370	0	0

bm = Biomart()

var_names = adata.var_names.to_list()

df = pd.DataFrame()

for i in range(1,int(len(var_names)/400) +2):
    subl = []
    if i*400>=len(var_names):
        subl = var_names[(i-1)*400:len(var_names)]
    else:
        subl = var_names[(i-1)*400:i*400]

    queries ={'ensembl_gene_id': subl } # need to be a dict object
    results = bm.query(dataset='hsapiens_gene_ensembl',
                       attributes=['ensembl_gene_id', 'external_gene_name'],
                       filters=queries)
    if df.shape[0]==0:
        df = results
    else:
        df = pd.concat([df, results], axis=0)

df.shape

(32926, 2)

df.index = df["ensembl_gene_id"]
df = df.drop_duplicates(keep="first")

not_conver_index = df[df["external_gene_name"]!= df["external_gene_name"]].index.to_list()
df.loc[not_conver_index,"external_gene_name"] = not_conver_index

df.shape

(32844, 2)

adata.var["external_gene_name"] = adata.var.index
have_symbol = adata.var.index[adata.var.index.isin(df.index)]

adata.var.loc[have_symbol, "external_gene_name"] = df.loc[have_symbol, "external_gene_name"]

adata.var.index = adata.var["external_gene_name"]

adata.var_names

Index(['MIR1302-2HG', 'FAM138A', 'OR4F5', 'ENSG00000238009', 'ENSG00000239945',
       'ENSG00000239906', 'ENSG00000241599', 'DDX11L17', 'WASH9P',
       'ENSG00000228463',
       ...
       'ENSG00000277196', 'ENSG00000277630', 'ENSG00000278384',
       'ENSG00000278633', 'ENSG00000276345', 'ENSG00000277856',
       'ENSG00000275063', 'ENSG00000271254', 'ENSG00000277475',
       'ENSG00000268674'],
      dtype='object', name='external_gene_name', length=32922)

adata.var_names_make_unique()

index name can not be dup to the column name

adata.var.index.name = "index_name"

adata.write("kidney_tumors_convert_symbol.h5ad")