[Bioc-sig-seq] Paired end reads -tools?

[Nicolas Delhomme]

Hi Paul,

First to clarify, are you talking about paired-end (insert size of  
around 200bp) or mate-pair (insert size of around 2-3kb) sequencing?

I've been doing some analysis of mate pair data in R (for a de-novo  
assembly). The synopsis of what I've done is (I can elaborate if you  
think what I've done is of interest for you):

1) load the aligned reads

2) filter them for N containing reads, reads aligned to non-ref  
chromosome, etc

3) identify the uniquely mapped mate pair, i.e. each mate map only  
once in the genome and has a single mate (computationally performed  
using their ids)

4) convert the mate reads into intervals stored into a RangedData object

5) estimate the insert size distribution and further processing  
(assembly in my case)

In your case, if you're using mate pairs (I hardly see how to do that  
with "standard" paired reads), you could have a window sliding along  
your chromosome and estimating whether the insert size in your cancer  
sample is different from the one in the normal sample.

Best,

Nico



---------------------------------------------------------------
Nicolas Delhomme

High Throughput Functional Genomics Center

European Molecular Biology Laboratory

Tel: +49 6221 387 8426
Email: nicolas.delhomme at embl.de
Meyerhofstrasse 1 - Postfach 10.2209
69102 Heidelberg, Germany
---------------------------------------------------------------



On 26 Oct 2009, at 02:31, Paul Leo wrote:

> I have some low coverage paired end data. I would like to plot the
> difference between the paired end reads.
>
> I've used Shortread, IRANGES etc some but have not seen explicit use  
> of
> paired end data ... I'm not quite sure how that data is handled in
> IRANGES as a RangedData object?
>
> Would be grateful any pointers or ideas on what others have used to
> analyse this data.... looking for structural varions in matched
> cancer/normal based on  paired in read  distance( data is very low
> coverage so I think this is all I can hope for at the moment...)
>
> Thanks
> Paul
>
> _______________________________________________
> Bioc-sig-sequencing mailing list
> Bioc-sig-sequencing at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioc-sig-sequencing