Many Broken Pairs

[Flo1000]

Hi everybody,
sometimes I have samples, where every other quality parameter I can think of looks fine, but I have 20% or even more broken pairs i.e. mate pairs are mapped far away from each other. What might be the technincal or biological reason for this?

[GenoMax]

Are you referring to paired-end reads? Mate-pair sequencing means something different and mate-pair reads are supposed to map far away.

If the mapping looks good for those reads (and there is support from multiple reads that do the same thing) then those events represent real splicing that must exist in your samples.

[Flo1000]

Yes, I'm sorry, I meant paired-end. Okay, but this is still very strange to me. How could non-tumor samples acquire this many splicing events?

[nucacidhunter]

If you are referring to DNAseq and pairs map randomly to the same (unexpected distance) or different chromosomes, they could be PCR fusion products.

[Flo1000]

>If you are referring to DNAseq and pairs map randomly to the same (unexpected distance) or different chromosomes, they could be PCR fusion products.

Yes! This is exactly the case. How can we prevent PCR fusion products? Are there some information available online?

[nucacidhunter]

PCR free library prep or minimising cycles.

[Flo1000]

Thank you!

[Flo1000]

I am still wondering what "PCR fusion products" really are and why they occur...can't find any useful information. Is there a paper, manual or something about it?

[Markiyan]

Check short fragments cleanup amd ligation conditions.

So those "broken" paired end reads should be called reads from chimeric templates (or inconsistent read pairs).

Usually they are caused by too high molar ratio of the template DNA to adapters concentration. The solution is similar to ones used for chimeric reads in the case of shotgun subcloning using PUC or similar vector:

1. Make sure you don't have any fragments under 150 bp of size (check your's ampure washes ratio and bioanalyser trace). 1ng of 50bp fragments is equimolar to 10mn of 500bp fragments, also those small bits ligate really efficiently.
2. If your fragmentation and small fragments removal works well, than try using more adapters and less DNA, also try increasing ligation reaction volume.
3. Avoid excessive freeze/thaw cycles on the kit reagents and adapters, since they would remove last few bases from the adapter, so they can self-ligate.

[Flo1000]

Thank you. We will check the fragment size distribution. But I still don't understand the process complety. Do you mean that during the PCR, different fragments from different chromosomes are ligated to each other?

[Markiyan]

Any DNA fragments present in the solution can get ligated together...

During a ligation step (circulisation / sequencing adapters ligation) for nextera matepair) any DNA fragments cpresent in the solution an get ligated together.

In normal reaction conditions (sufficient dillution and lack of tiny fragments) the ends of a circle are mole likely to get ligated together, than form a chimera with another insert, but if the concentrations is too high or there are quite a few tiny fragments (which have high ligation efficiency) than they can form chimeric templates.