Convert Entrez gene ids and HUGO symbols, whose genome assembly is unknown, to GRCh38 Ensembl gene ids

Question

I want to use a file available on a database called DepMap (https://depmap.org/portal/download/). This file called CRISPR_gene_effect.csv contains gene ids in the format HUGO (Entrez) eg. TP53 (7157), without having any complementary information regarding which genome assembly (GRCh37/38) the Entrez ids (and also HUGO symbols) belong to.

I want to convert these gene ids in the HUGO (Entrez) format to GRCh38 Ensembl ids (specifically to Ensembl release 100), so that I can map the data to Ensembl's annotations. I would like to get suggestions on this.

I feel that figuring out the genome assembly version (GRCh37/38) of the Entrez ids would be critical in this conversion. Although, I am not aware if Entrez ids change from one genome assembly to other.

Answer 1

Hugo names don't really have an associated assembly. They are just the official name of the gene in human. You will find them in all assemblies (except if they hadn't been annotated/identified in one of the older ones). Entrez entries are also not tied to a specific genome assembly. They usually have mappings to multiple ones. For example, Hugo gene name A1BG, with Entrez ID 1 has:

Now, if you want to get Ensembl gene IDs for a list of Hugo gene names, you can use BioMart to convert. First, get the list of target IDs. You can use either Hugo or Entrez. I found the Entrez simpler to parse out, so I went for those (note that this assumes GNU grep):

grep -oP '(?<=\()\d+(?=\))' CRISPR_gene_effect.csv > entrez.ids

I then went to BioMart and selected "Ensembl genes 104" (note that I could only get Ensembl version 104, I assume you're not adamant about getting version 100 given that it doesn't seem to be current) and "Human genes GRCh38.p13" :

Next, I set the filters to "NCBI gene (formerly Entrezgene) ID(s)" and uploaded my file:

And, finally, selected the Attributes I want to see, namely Ensembl gene ID, HGNC (Hugo) symbol and Entrez ID so I can map everything to each other:

I then clicked on "Results" and got:

[

You can now download this to a file and have your mapping. You might even be able to use the one I created: https://www.ensembl.org/biomart/martview/352aae8e0b5c774cbbee73d67634d577.