[Lovro]

Thank you for all the info and this nice summary.
I was wondering if you guys have any advice on eliminating large fragments. The step is included in TruSeq DNA PCR-Free library prep from Illumina. They dilute 100ul of SPB with 60ul water for 350bp insert and with 100ul of water for 550bp insert size and then use supernatant in the next step.
I'm trying to prepare 1000bp insert size library and am wondering if I should perform this step at all since according to gel, there aren't any larger fragments in the samples (DNA was sheared with Covaris)
Are large fragments problematic downstream in the protocol, including sequencing on MySeq?
If I will perform this step, maybe I should dillute SPB with water 1:3 and this will leave 1000-1500bp fragments in the supernatant? This is only speculation, so any advice from experience would be appreciated.

On the other hand, since I will later enrich the library with custom IDT probes, the 130bp adapter ligates probably would not be a problem since they will not be pulled ?

There is also no PCR step before enrichment, so biased replication of the shorter fragments shouldn't be a problem.

Would you recommend skiping both cleanup steps in order to preserve the sample?
Your thoughts ?

Thank you,
Lovro

[nucacidhunter]

To prep a library with fragments over 1kb, you need to shear DNA to 1kb with mini-tube and do one sided size selection using 0.45-0.48x AMPure beads. Larger fragments will not interfere with sequencing.

After ligation excessive adapters need to be cleaned up otherwise they will persist throughout the capture and might saturate blockers resulting in inefficient and off target pull down.

[Lovro]

Thank you!
Almost forgot about the blockers used in enrichment.

By one sided you mean I have to get rid only of the shorter fragments? With 0.45-0.48x AMPure beads you suggest I keep the beads or the supernatant? In the official protocol they use such bead concentration to bind large (I guess over 600bp) fragments and then use supernatant (remove large fragments).
Keeping the beads would then preserve everything above cca 600bp for further steps.
However, I am concerned how much DNA will I loose in the process ??
The reason I'm worried is because with enrichment step, most of the DNA will be discarded.

I would loose less DNA with, say, 0.7X spb. However this would preserve shorter fragments (still above 300 bp probably?) as well. Is there a reason I should be worried about this in regard to sequencing efficiency. In essence I'm interested in sequencing shorter inserts as much as I am in longer. I know the shorter fragments will cluster more, but I don't know to what extent. As long as I get 10% of 1000bp inserts it should be fine.
But I don't want to risk more sequencing errors or data loss because of uneven insert sizes.
So i guess my next question would be: "does uneven insert size hinder the sequencing process? "

Thank you!

[nucacidhunter]

You need to keep the beads and after washes elute large fragments bound to them.

The amont of DNA lost will depend on size distribution of sheared DNA. You will loose fragments below cut off and a bit above it.

To sequence 10% of 1kb fragments libray size roughly should be 800bp-1.5 kb.

I do not know what you mean by uneven because regardless of size cut off you still will have bell shaped size distribution.

[Lovro]

I was under impression that bead size selection is one sided and different concentrations of SPB bind DNA from certain length upwards. Setting the threshold at lower lengths would therefore result in wider insert size distribution.
On the other hand, if certain SPB concentration binds DNA with fragment lengths inside a given interval, the wideness of insert size distribution would remain constant.
I've read somewhere that 1:1 ratio of spb and sample binds everything of 200bp upwards. Therefore I thought the size selection is (predominantly) one sided (??is it).

Although I think that the wideness of the distribution should not be a problem sequencing-wise (there will not be more sequencing errors), I am aware of clustering being short fragment biased.

On one hand, I want to have large insert size. On the other I want to preserve sample. Too large fragment size will waste sample also because I plan to spike the library into routine runs with shorter fragments.

Anyhow, I will try with one sided size selection like you suggested, but with 0.55 AMPure beads. I'll report the results

tnx

[nucacidhunter]

Binding DNA to beads is one sided but size selection can be left-sided, right-sided or double-sided. Following link has good explanation and discussion on the mechanism.

http://core-genomics.blogspot.com.au/2012/04/how-do-spri-beads-work.html

There are different brands of beads sold for clean up or size selection but their cut off and recovery efficiency varies, so you would need to do some trial to find the optimum bead/DNA ratio for your application.

If you spike large fragment library into short fragment one you probably will get small number of sequence reads from large fragments.

0.55x will cut somewhere around 600 bp depending on bead brand.


Edit: Ligation reactions containing PEG will affect size cut off.

[Lovro]

Thank you for the link.

I will use the beads that come with Nextera kits:



Are there any specificities on them I should know about ?

In regard to the edit: More peg means shorter fragments will be included ? Do you happen to know, if the effect will be large with Illumina TruSeq PCR-free kit?


Thank you!

[nucacidhunter]

The beads in the picture are from TruSeq kit. They are SPRIselect or AMPure XP from Beckman. Nextera kits do not contain any beads although Nextera XT comes with library normalization beads.

TruSeq PCR free ligation reaction does not seem to have high concentration of PEG so the effect will be minimal and anyway there is a second clean up as well.

[Lovro]

They are from Nextera exome kit, not generic nextera.
http://www.illumina.com/products/nex...xome-kits.html

tnx for all the info!