A question about annotation

[xiaowenchenjax]

Hi authors,
my scATAC-seq data were aligned to the reference genome "gencode vM17", I looked through the previous discussions about it, I seems like I need annotate scATAC-seq data by this annotation file. I can provide TSS position when running TSSEnrichment() function, when calculating gene activity, i need get the coordinates of genes using the following codes:
gtf <- rtracklayer::import('genes.gtf')
gene.coords <- gtf[gtf$type == 'gene']
seqlevelsStyle(gene.coords) <- 'UCSC'
gene.coords <- keepStandardChromosomes(gene.coords, pruning.mode = 'coarse')

My question is why we need sequence level is UCSC? Do I need to run this script when i used the reference from gencode? if not, what sequence level I need to assign?

Thank you.

[xiaowenchenjax]

the sequence level is UCSC that means chromosomes like this chr1, chr2...

[timoast]

You need to set the seqlevelsStyle for you gene annotations to match whatever genome the data were mapped to. This will not always be UCSC, that is just given as an example on the Signac website.

[xiaowenchenjax]

Thanks, based on the following the script
annotations <- GetGRangesFromEnsDb(ensdb = EnsDb.Mmusculus.v79)
mm10.tss<-GetTSSPositions(annotations)
Can I get the correct TSS position? seems like annotations contain the information of all the exon,gap , utr , cds for one gene.
What is gap? Based on discussion "GeneActivity function and TSS location #508", you collapse each gene down to the longest transcript and take the start of that transcript as the TSS for the gene.
when I check the script, for one gene, i found TSS is the min start value of exon, gap, utr, cds if strand = +, otherwise, TSS is max end of all exon, gap, utr, cds. but transcripts is not composed of exon, gap, utr, cds randomly.
Is this method of calculating TSS correct? or I miss something?