What fragment size do you use for your Illumina libraries?

[kmcarr]

Illumina recommends insert sizes of ~200bp for SE libraries and ~300bp for PE libraries but clearly indicates in their literature that you can go larger (500-600bp). I'd like to ask the NextGen community a couple of questions:

1. What insert size do you routinely use for your SE & PE libraries?

2. Have you noticed any relationship between overall sequence quality and library fragment size? (Actual data correlating these would be terrific.)

Thanks.

[kainsteven]

For RNA-Seq applications, we have seen better results by excising material corresponding to the peak of the bioanalyzer trace for library preparation. In most cases, this corresponds to ~ 250 – 350 bp in length. We have used this range for both SR and PE libraries, but only with short read lengths of 32-36 bp. For longer read lengths and PE sequencing, I would imagine you would want to lean towards the longer end of this range.

Best,
Steve

[greigite]

I don't have any data on this myself, but this paper reports the successful use of a 600 bp library, which reduced overall coverage variability: Harismendy & Frazer 2009

[maureen]

Solexa recommends 200-250 bp and a tight range, but I definitely know of successful routine sequencing of libraries 300-700 bp. In fact my own library is 300-600 bp.

[der_eiskern]

illumina may recommend smaller library sizes but i'm pretty sure there are people using 1kb or 2kb insert sizes in parallel with the 200-700 bp sizes for structural variant discovery.

[edge]

Hi kainsteven,

I have an Illumina RNA-seq, pair-end read, 2x90bp....
However, I not really sure about the fragment size/insert size
It might affect my assembly result.

Based on your experience, do you have any idea regarding the fragment size of Illumina pair-end RNA-seq data if my read length is 90nt.

Many thanks for advice.

For RNA-Seq applications, we have seen better results by excising material corresponding to the peak of the bioanalyzer trace for library preparation. In most cases, this corresponds to ~ 250 – 350 bp in length. We have used this range for both SR and PE libraries, but only with short read lengths of 32-36 bp. For longer read lengths and PE sequencing, I would imagine you would want to lean towards the longer end of this range.

Best,
Steve

[RPiddock]

Our libraries team generally shear for about 500bp with Illumina TruSeq libraries.

[edge]

Thanks for replying
Is there got any other way to inferred insert size by undergo some statistic analysis?
Thanks.

Our libraries team generally shear for about 500bp with Illumina TruSeq libraries.

[tony19871929]

question about illumina seq.

hello every one , i really want to know the different Illumina standard (300bp insert) from overlapping 180bp.? can you help me ,thank you

[kainsteven]

Thank you for the question. As mechanical or enzymatic fragmentation methods will produce a distribution of fragment sizes we typically choose a median size of 350 to 400 bp for 2 x 90 bp reads. There is no harm is going to this length, and it reduces the likelihood of any forward and reverse sequencing overlaps.

Steve

[apeterson]

illumina may recommend smaller library sizes but i'm pretty sure there are people using 1kb or 2kb insert sizes in parallel with the 200-700 bp sizes for structural variant discovery.

Does anyone have any more information on making 1kb insert size libraries?

[cement_head]

Hi kainsteven,

I have an Illumina RNA-seq, pair-end read, 2x90bp....
However, I not really sure about the fragment size/insert size
It might affect my assembly result.

Based on your experience, do you have any idea regarding the fragment size of Illumina pair-end RNA-seq data if my read length is 90nt.

Many thanks for advice.

It should not affect your assembly, IF you have enough (ideal) coverage. What will affect your assembly is if you have overlapp in your SBS. For example, if your fragment length was 300 bp and your paired-end reads were 200 bp you will have a problem because the reads actually have 100 bp overlap and the assembler will have difficulty discerning whether or not it's an overlap or two different reads.

What you need to calculate, or consider, is how much of the end of the fragment that you need to sequence in order to identify it as unique. In a haploid organism that answer is not very much. In a diploid organism that answer is more complicated. You have to consider variants (like alleles) and how you will be able to assign reads to a given gene. Then you might consider SBS to within a few bases of each other (e.g. paired-end sequencing of 140 bases of SBS on a 300 bp fragment) to cover most of the fragment.

How did you prepare your library? If you used Nextera, you have sizes that range from 300 to 1400 bp. If you used disruption, you should have a much narrower distribution of sizes.

Hope this helps,
Andor

[chiaraf]

Hi!
I routinely sequence libreries with a mean fragment size of 450 bp for a 100PE for DNA and libreries witha mean fragment size of 350-400 for a 100PE for RNA seq.

[pc2009open]

Hi everyone,

I am also interested in 1Kb-1.5Kb library. Do you know any lab that has the expertise?

Thanks a lot.

Does anyone have any more information on making 1kb insert size libraries?