Mixing libraries with different insert lengths in a lane

[MesutOezil]

For sequencing on illumina HiSeq, I am now making DNA libraries with different insert lengths (original lengths of just sheared DNA fragments peaked at 300, 450, and 600).

In sequencing, do you recommend not to mix two or three of them (with different insert lengths) in a single lane? Will there be any unfavorable effect caused by the different insert lengths in a lane? Maybe libraries with shorter inserts are more frequently sequenced?

[simonandrews]

In general you get smaller clusters and stronger signals from shorter libraries so you might get some bias towards the shorter libraries in your lane. I don't think you're likely to see any bias other than coverage, and if it was my experiment I'd probably be willing to risk it to see what I got.

[MesutOezil]

Thanks, simonandrews.

I see. Then, the next question is about PhiX which I believe has shorter inserts. With libraries with longer inserts, does it make sense to include a smaller proportion of PhiX than usual (=than that in other runs including libraries with similar lengths of inserts).

[HeinKey]

We are sequencing libraries with mixed fragment lengths for years now. Works fine in our hands. Images are looking more messy, but read quality is ok. Our fragments vary between 200-700 nts in single libraries. We spike 1% Phi-X and sequence on the HiSeq.
Hope this helps.

[IDDan]

HeinKey is correct.

Fluidigm has a HighThroughput amplicon application for Illumina that gives you just this (amplicons of varying length) and it works just fine for us...worked on 454 also.

That being said, we haven't designed/sequenced amplicons over ~800bp and keep them over ~300bp so we don't lose too many (ampure size-selects so smaller amplicons are lost in higher proportions) during bead cleanup.