[Bioc-devel] GenomicFeatures Introns

[Dario Strbenac]

What do you think about RNA-seq reads over splice junctions ? Getting just the GRanges of the gap coordinates of each gapped read (from a GappedAlignments object), and overlapping them the gene introns list with type = "equal" using something like findOverlaps, would give you, for each possible splice junction of a gene, the number of reads that spliced across it ?

---- Original message ----
>Date: Wed, 11 May 2011 10:04:26 -0700
>From: Marc Carlson <mcarlson at fhcrc.org>  
>Subject: Re: GenomicFeatures Introns  
>To: D.Strbenac at garvan.org.au
>
>Hi Dario,
>
>That is an interesting question with an interesting answer which I will 
>attempt to describe.
>
>Our reasoning was that since introns are basically defined as the gaps 
>between exons within a transcript, we were not sure whether it would 
>ever really make sense to group them in another way.  Basically I can 
>easily imagine why someone would want to count all the exons for a gene 
>or even why they might reduce something like that and then infer the 
>resulting gaps as universally intronic sequence (for that gene).  But 
>that set of operations is NOT the same thing you would get if you 
>counted starting with a potential intronsBy() method.  In fact, if you 
>had started with a naive intronsBy() method you might have horribly 
>miscounted since you would have all sorts of exonic regions contained in 
>some of your introns...  :(
>
>But perhaps you have another idea of how to sort this out?  If you have 
>a compelling use case (and a way to do it that makes sense), please tell 
>us about it on the biod-devel mailing list.
>
>
>   Marc
>
>
>On 05/11/2011 04:00 AM, Dario Strbenac wrote:
>> Hello,
>>
>> I have a short design question. In a wiki page
>>
>> http://wiki.fhcrc.org/bioc/Annotation_Helpers_Spec
>>
>> It has a intronsBy(txdb, by) function described, but in GenomicFeatures, there's a intronsByTranscript function. Is there a reason why the Wiki intronsBy options were dropped (which is more like how exonsBy currently works) ? I didn't ask on the list, in case there was an obvious reason.
>>
>> - Dario.
>


--------------------------------------
Dario Strbenac
Research Assistant
Cancer Epigenetics
Garvan Institute of Medical Research
Darlinghurst NSW 2010
Australia