Hi! I have a 10X scRNAseq data set that contains pig and human cells. I am trying to generate a hybrid human/pig reference for CellRanger as described by Cheung et al Nature Comm 2024
My approach is to download the ENSEMBL pig and human FASTA/GTF files. In order to ensure fully unique gene names, I wanted to add a prefix 'HUMAN' and 'PIG' to each. However, I am not sure where to add the prefix in both files bc their nomenclature seems to be different. For example, below, I could make it "HUMAN_ENST00000390473.1" and "HUMAN_ENSG00000142611" but I'm not sure how CellRanger could map them together.
Any ideas? Please let me know if I am misunderstanding or approaching this wrong.
#FASTA EXAMPLE:
ENST00000390473.1 cdna chromosome:GRCh38:14:22450089:22450139:1 gene:ENSG00000211825.1 gene_biotype:TR_J_gene transcript_biotype:TR_J_gene gene_symbol:TRDJ1 description:T cell receptor delta joining 1 [Source:HGNC Symbol;Acc:HGNC:12257]
ACACCGATAAACTCATCTTTGGAAAAGGAACCCGTGTGACTGTGGAACCAA
#GTF EXAMPLE:
1 ensembl_havana gene 3069168 3438621 . + . gene_id "ENSG00000142611"; gene_version "17"; gene_name "PRDM16"; gene_source "ensembl_havana"; gene_biotype "protein_coding";
1 answer
A prefix for ensembl genes is not necessary. ENSG is strictly human; other species have different prefixes (ENSSSCG for the pig reference).
You only need to:
(1) Concatenate the fasta files, adding a prefix or suffix to the contig names
(2) Concatenate the .gtf files, adding the same prefix or suffix to contig names
(3) Follow the CellRanger mkref workflow.
Log in to answer this question.
Looks like 10x has a recommendation to create a reference with multiple species: https://www.10xgenomics.com/support/software/cell-ranger/latest/analysis/inputs/cr-3p-references#multiple-species-4f40e4