Illumina HiSeq Inconsistent Reads

[jeannelouise]

If you pool 6 samples into one tube and put them into one lane of a flow cell, the idea is to get approximately the same number of reads per sample, is that correct? If you don't, what might be the cause of that?

[Brian Bushnell]

If the samples had different volumes, or different concentrations, or different efficiencies in some way (for example, the fraction of the DNA that correctly ligated to adapters), or different insert sizes, then their relative abundance in the output will be affected. Those are some of the most common reasons.

[jeannelouise]

Mi Seq before Hi Seq

Do most users run their samples on the Mi Seq before the Hi Seq with the hope of ruling out problems quicker and with less expense? Would this be recommended to help with a problem of inconsistent number of reads per sample in the same flow cell?

[GenoMax]

Do most users run their samples on the Mi Seq before the Hi Seq with the hope of ruling out problems quicker and with less expense? Would this be recommended to help with a problem of inconsistent number of reads per sample in the same flow cell?

Not generally. MiSeq runs do have a cost.

We have done MiSeq runs for projects that pool tens of samples together to make sure that the pool is balanced/to decide on how to re-pool samples when needed.

Can you clarify what you mean by "inconsistent number of reads per sample in the same flow cell"? Is that for repeat runs or just a single run for an unbalanced pool?

[jeannelouise]

By "inconsistent number of reads per sample in the same flow cell", I am trying to ask if you pool the same volume of 6 libraries with the same concentration into a single tube and put them on one lane of a flow cell, I expect to get the same number of reads for each library. I am not getting consistent number of reads per library, resulting in customers not happy and then I must rerun the libraries at my expense. I was wondering if running the libraries on the Mi Seq first would show the inconsistencies quicker than running them first on the Hi Seq which is a longer run and more costly?

[atcghelix]

How are you quantifying the concentrations of your samples prior to pooling?

[jeannelouise]

I am quantifying using Qubit and Agilent Bioanalyzer.

[atcghelix]

Do you have access to a qPCR machine? I'd suggest trying that out to quantify your samples.

http://seqanswers.com/forums/showthread.php?t=6362

[sarvidsson]

What kind of libraries are you sequencing? E.g. PCR-free libraries are very difficult to quantify correctly.

[jeannelouise]

We do PCR-free libraries, as well as others. That is part of the problem. We are a core facility and many times the libraries are prepared by the customer and we do not always know how they have been prepared.

[GenoMax]

This may come down to how you manage expectations of your customers. You can tell them that you are doing your best to balance the pools with qubit/bioanalyzer. If they absolutely need to have comparable reads then they would need to pay for a MiSeq run. You will find that most people would be ok with what they get by your method.

That said, concentrations is just one factor here. Depending on the size of the inserts, quality of the libraries you are going to get some samples that will cluster better than others.

Are you mixing samples from different submitters in one lane?

[jeannelouise]

Illumina HiSeq Inconsistent Reads

I am troubleshooting the problem of inconsistent read numbers on Illumina Hiseq. To what extent does the type of DNA affect read numbers? Does the size of the genome make a difference? Do GC rich libraries make a difference? Does this result in bias where certain types of fragments are overlooked?

[jeannelouise]

Illumina HiSeq Inconsistent Reads

When trying to solve the problem of inconsistent reads on Illumina HiSeq's, does overclustering affect the read counts? Can DNA be quantified that does not form clusters, creating lower read counts, such as a DNA fragment with no adapter? Has anyone had experience with adapter ligation and barcode indexing affecting read counts?

[jeannelouise]

Illumina HiSeq Inconsistent Reads

Still working on trying to solve the inconsistent read problem, what types of problems have been seen regarding library prep in relation to inconsistent reads? Areas I am questioning are concentrations, volumes, insert sizes, fragment sizes. Also, what about fragmentation methods, NaOH denaturation and library contamination?

[jeannelouise]

Illumina HiSeq Inconsistent Reads

When quantifying and normalizing library preps, can anyone give me their thoughts on which methods of quantifying and normalizing give best results and most consistent read numbers? I am interested in your thoughts on Agilent, Fragment Analyzer, Qubit, Picogreen, qPCR, and nanodrop? Has anyone had experience with over- and under-quantification as a problem leading to inconsistent read counts?