splitting of single-end reads into paired-end reads: how is it done?

[emariella]

Hi Daniel,

I have just started to use FusionCatcher to analyze some single-end RNA-seq data derived from tumor samples and for the moment the results are in line with my expectations. However, although I know that single-end reads that are longer than 130 bp are automatically splitted into paired-end reads (reads are 150 bp long in my case), I cannot find a description of this step neither in the manual nor in the preprint. If you could provide me some details about how the splitting is done, it would be helpful to me for better result interpretation and visualization.

Thanks in advance!!

Elisa

[ndaniel]

Hello Elisa,

basically, one single end reads is split into several/many paired-end reads using the fragment_fastq.py. The longer single-end read is the more paired-ends reads will be generated. This will have the side-effect that a candidate fusion junction might reported as being supported by several paired-ends (which actual come from one single-end read)

Here is a graphical representation:

sedr    ---------------------------------------------------------------------------------------
per1    -------------------------..-------------------------
per2    -------------------------..................-------------------------
per3    -------------------------.....................................-------------------------

where sedr (single-end read) is split into 3 paired-ends reads that are per1, per2, and per3.

[nihilee]

@ndaniel dear developer, I am curious about why break raw reads to fragments to do latter align job, is it helpful in finding fusion? I thought this step may increase a lot working time.