MiSeq tolerated library size distribution

[ES5063]

Our lab is trying some new library generation protocols where we are size selecting a much broader range of libraries than we have run previously, a 200 bp span as opposed to a 20 bp span. When we tried sequencing this pool on MiSeq, v2 300 cycle kit single-end, the run gave an out of focus error about halfway through. It seems like this corresponds with something seen in our bioanalyzer traces, that the libraries are not evenly distributed across the size range but heavier on the smaller side. My thought was that once all these smaller libraries were sequenced, there wasn't enough larger libraries left for the MiSeq to maintain good focus? PhiX was included at 5%.

Is there a maximum tolerated spread of library sizes in a MiSeq sequencing run? Would simply spiking in a higher percent PhiX resolve the issue (10 or 25%)?

We are planning to change to paired-end sequencing moving forward as another method to hopefully help with this issue.

[GenoMax]

When you say 20 bp span does that mean you have just a 20 bp insert?

[ES5063]

The span that I cut when size selecting on a gel. Prior libraries have been ~140-160 bp (total library size). The new ones are more like ~140-340 bp.

[GenoMax]

Smaller inserts do tend to cluster more efficiently on flowcells. Can you tell us what the cluster density was? You may need to deliberately underload the run to keep cluster density in check and/or add more phiX.

It would be useful to have Illumina tech support look at the run metrics remotely and see what they say.

[ES5063]

Cluster density was higher than I anticipated, loaded 6 pM and density was 1300K/mm2. I'll keep that in mind to load less than I typically would for future runs.

Yup, hoping to have tech support weigh in as well. Thank you for your quick responses!

[jdk787]

It's likely that you are just sequencing off of the end of the library, since your read length is longer than most of your inserts with adapter on the end.

Why use 300 cycles when you largest library size is 340bp, which means your largest insert size is at most ~220bp?

More PhiX will help and a 2X150 run would probably produce better results, but you'll still be sequencing off the end of any library that's shorter than 200bp so you might just want to use a 2X75.

[ES5063]

Honestly using the 300 cycles was a learning curve. Previously only needed 50 cycles kits so was initially told you put down the number of cycles of the kit in the sample sheet. The idea of sequencing through the entire insert didn't even occur to me. I hadn't quite pieced together until after the run failed that the you don't read through the entire library, only the insert (Still consider myself a newbie here). That being said we were able to recover some of the data from the failed run since it read straight through the index and our bioinformatician could pull out that information to index the sequences.