How To Get Non-Overlapping Sequences Of Introns, Utrs, Flanking Regions For A Gene?

Question

[liux.bio]

Hello,biostars. I am using bioconductor to get sequences of introns, UTRS, flanking regions for genes. I have a set of genes, I want gene-centered sequences of introns, UTRS, flanking regions. I am using packages such as Biostrings, GenomicFeatures. I gain sequences of UTRS, flanking regions, introns for transcripts of a gene. But I don't know how to delete the overlapping sequences(repetition sequences?). Maybe it's so naive. Any suggestions? Many thanks and Happy New year!

[Devon Ryan]

Presuming you have a GRangesList containing exons with transcript information split by gene, have you tried reduce()?

[liux.bio]

Yes,I tried. I read the help but I can't understand it. For a GRangesList split by transcripts, it deletes all the overlapping between the transcripts and in the transcripts,right? I will read the help carefully.Thanks!

[Devon Ryan]

Ah, you have things split by transcript. You might instead split things by gene and then use reduce (this can be easily done if you directly import the GFF/GTF file into GenomicRanges, I don't use GenomicFeatures so I can't say how things work there). If you have things split by gene, then what reduce() will do is merge overlapping exons between transcripts to create a "union gene model" (i.e., what you would get if you collapsed all of the transcripts together), which sounds like what you want. That way, there are no repeated regions.

Edit: I'll add that you can unlist() a GRangesList that's split by transcript and then split() it by gene.