I am using several cellranger tools/pipelines to analyze some data, specifically, mkgtf, mkref, count, and aggr. I am using mkgtf and mkref to include a PI-synthesized protein to my reference. I have finished the analysis but am not seeing the protein in my output, so I am going back to the start to verify the process.

Unlike the documentation (https://support.10xgenomics.com/single-cell-gene-expression/software/pipelines/latest/advanced/references), my GTF file was not downloaded from ENSEMBL or UCSC, but NCBI (https://www.ncbi.nlm.nih.gov/nuccore/AY678269.1?report=genbank). This GFF file does not have the same key nomenclature as seen in ENSEMBL or USCS.

I did use the mkgtf tool, but the created file is identical to the input.

cellranger mkgtf custom_ref/AY678269_genes.gtf reference_data/AY678269_mkgtf_genes.gtf

I went ahead and used cat to add the fasta and gtf file from the synthesized protein to the respective reference fasta and gene files, then ran mkref.

The s.protein is in the fasta file:

>lcl|AY678269.1_cds_AAV52169.1_1 [protein=tandem-dimer red fluorescent protein] [protein_id=AAV52169.1] [location=1..1431] [gbkey=CDS] 

as well as the gene list

##sequence-region AY678269.1 1 1431
##species https://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?id=32630
AY678269.1      Genbank region  1       1431    .       +       .
AY678269.1      Genbank CDS     1       1431    .       +       0

However this protein is not found in the output...surely I must have made an error. Does anyone have any ideas about what I could be doing incorrectly?

Do I need to change the entire nomenclature of the GFF keys to match what I see from ENSEMBL (e.g., gene_id ""; gene_version ""; transcript_id, etc...)?

Any assistance is greatly appreciated.


Regarding your comment:

This GFF file does not have the same key nomenclature as seen in ENSEMBL or USCS

The link you have provided corresponds to an entry in Genbank format and not to GFF format. I believe you need a file in GTF format in particular.

Nevertheless, given the fact that you have manually added the corresponding entries to the reference fasta and gtf files prior to mkref, you should have seen your gene of interest in the output. I would double check one of the original fasta headers and original gtf entries in the reference file and match these for the new entry.

  • $\begingroup$ Thank you for the response. I figured out what the problem was, I was given an incorrect accession number! However, after finding & reading through this: support.10xgenomics.com/single-cell-gene-expression/software/…, I can see how my approach may have been problematic anyway because of the NCBI keys. Fixed it & I'm creating the custom reference. FYI, the link does go to a genbank file, but when downloading the information, you have the option of choosing the format (fasta, genbank, gff), but you are absolutely correct about the GTF vs GFF. Danke! $\endgroup$
    – cer
    Mar 3 '20 at 14:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.