Custom annotation - help to bypass/change it for GO, TF, etc-ATAC-related analysis

[NadiaEdel]

Dear Signac team,

Thank you so much for this amazing tool and for keeping up with all of our questions.

I have seen similar issues in other posts, but I could not apply their fixes to mine.

I have a custom-annotated genome (not my choice, and I cannot get it changed to the standard UCSC/NCBI one). This is preventing me from any deep analysis on the ATAC data (no JASPAR run for TF binding and motifs, not able to move forward with LinkPeaks() function, etc).

The chromosomes are labelled as "NC_" instead of "NC." (NCBI style) or the UCSC style. I was wondering if and how can I change this in my Seurat object (I have multiome samples, RNA+ATAC, I have already merged, integrated, etc...but if I can fix this annotation issue I can run it again).

I am working in R(studio) on WindowsOS, but could get my hands on Unix if needed.

I am struggling so much with this, I tried different things and could not fix it. I have the equivalencies between the annotation used and the NCBI and UCSC styles, I just do not know how to change the information in the seqinfo/names level...I have been reading and I understand it is more complicated than adding a column or something similar.

Apologies if the question is silly, but I have been stuck here for a long time now.

Thank you!

Kind regards,

Nadia

[timoast]

Just to clarify: the chromosome naming for analysis in Signac doesn't have to be in UCSC or NCBI format, it just needs to be consistent across the different things being used in your analysis. This would include the BSgenome object, gene annotations, fragment file, and peak matrix. It sounds like the simplest solution for you might be to rename the chromosome names in the BSgenome object and gene annotations so that they are in the format that you want, and then you should be able to proceed with the steps shown in the Signac vignettes.

[NadiaEdel]

Dear Tim and team,

Thank you so much for taking the time to answer my question.

Oh, I will try to do that, not sure how...I did not think it was possible to modifiy those files. Thank you for the suggestion!

But will I be able to perform motif analysis afterwards, co-accessible networks, TF footprinting and GO-analysis, or would that be a problem since there is always a "genome" or a "keyType" (for GO, for example, https://stuartlab.org/signac/articles/pbmc_vignette#non-linear-dimension-reduction-and-clustering).

For example, I was able to find DA peaks, generate the matrix from JASPAR2020, but then the "AddMotifs" was not possible because it kept telling me it did not recognise the genome.

Thank you again for your very kind help, and also I apologise for creating the post under "issue" instead of "discussion", thank you for changing that.

Kind regards,

Nadia

[timoast]

Yes, you should be able to run all of these steps. You don't need to modify any files, just change the chromosome naming in the objects in R. For example to change the chromosome names in the BSgenome object:

seqnames(BSgenome.obj) <-  gsub("NC_", "NC.", seqnames(BSgenome.obj))

(just an example, you will need to check how to rename them as needed for your project)