20% adapter get sequenced using MiSeq

[wendy_H]

Hi all,

I am new to MiSeq sequencing and it really frustrates me lately.

I had two MiSeq run recently and they all turned out to have about 20% adapter get sequenced. I've been using the TruSeq PCR-free DNA prep kit for library preparation. So the first time I followed the protocol for library prep ant it gave me 20% adapter sequenced. The second time I did an 8 cycle PCR enrichment for my library, followed by gel extraction of DNA fragments ranging from 500bp to 1500bp, hoping to remove all adapter dimers. However it gave me about 20% adapter sequence as well, which also leads to an uneven distribution of my libraries (percent reads pass filter for all samples ranging from 0.2% to 24%).

I wonder if anyone knows why this happens? We don't have bioanalyzer, so I only used qPCR for library quantification using KAPA kit. Also if anybody knows where I could find a commercial bioanalyzer? I would like to pay for them to run a couple of samples for me, in order to get a better MiSeq result.

Thank you!!!

[nucacidhunter]

I wonder which of following you are referring to:

1- whole read was adapter sequence for 20% of reads
2- 3' end of reads had adapter sequence which was 20% of output

The first case indicates incomplete removal of adapters after ligation while the second case would happen if read length were longer than insert size. If you have not use BA then how would you know the average size of libraries which is essential to calculate molar concentration from qPCR results.
PF% is very low which could be result of over clustering among other causes. You need to run sample in BA or a similar device to obtain correct average library size for optimum loading and clustering.

[wendy_H]

It's number 1. The most recent MiSeq was run using 2x300bp, in about 20% of the sequencing output, I get about 60bp read length followed by low quality nucleotides, which I think indicate number 1 happening.

For read length calculation, I ran the enriched library on DNA gel so I could get an estimate of average fragment length. I should correct myself, the cluster density (~1100K/mm2) and pass filter rate (~90%) for the MiSeq run was great. The problem I had was distribution of % reads. For some of my samples, they were sequenced for only 0.2% to 1% among all sequencing samples; for some other samples, they were sequenced 24%. Theoretically if I have 18 samples and I loaded them equally, I am expecting about 5.5% for each get sequenced.

[GenoMax]

Qubit for sample concentration along with bioanalyzer for size estimation is essential to get balanced pools. You will always find some variation in the actual number of reads you get for each sample (due to pipetting/conc estimation errors).

As you discovered yourself smaller inserts (adapter dimers qualify) cluster very efficiently compared to longer real inserts. So the only way to avoid that is to get rid of them by purification.

[wendy_H]

To GenoMax,

Unfortunately we don't have BA and can't afford to have it for now

To remove the smaller inserts, I did gel extraction (to remove fragment length less than 500bp), but it didn't seem to work

[GenoMax]

You could pay an additional charge and have your sequencing facility QC your samples on BA.

I am not on experimental side of things but bead purification would be better to remove primer dimers.

[wendy_H]

Yeah, so we don't have instrument for beads purification either... and we don't have a sequencing facility, so I did everything by myself in-house.

I wish I could pay sequencing company/facility just for BA usage.

[GenoMax]

I don't think you need an instrument for bead purification.

If you are in the Dallas area I am sure there will be local facilities that offer BA analysis as a "fee for service" option.

[wendy_H]

Cool! I will look into that. Thank you!