[kentawan]

Hi, Kentawan.

I am a beginner of NGS on Nextseq 500. I am primarily doing whole genome bisulfite sequencing with Nextseq 500. I saw your post on this forum and would like to learn from you.

1. based on what you said, you use V2 kit and got a very good cluster density. we used V2 but never get that high cluster density. Ours is about 140K. Can you share your experience on that? Appreciate!

2. We always had a big issue on Barcode reading, other metrics look very good. There are a high portion (35%) reads are reading as "GGGGGG", I think it is due to reading failure, but we do not know how to pin it down, we pooled two samples each time. Unfortunately, there was one time, we pooled with other 4 Chip-seq samples and still the same. Do you some idea how to fix it?

Thank you very much

Roger

Hi Roger,

We spiked in a good 20% freshly diluted PhiX library, as our library's first 6 bases are the same due to restriction enzyme processing.

As for your question 1, I'd like to know how do you do your library quantification? We used KAPA's library quantification kits. You might want to do it the old school way, clone the library into T-vector or something and Sanger sequence it to see if the P5 and P7 are sequence correct and in good shape. I suspect your P5 and P7s are not working properly.

as for question 2, again, would like to know if it's possible for you to Sanger sequence your library (let's say 50 clones would be good.). Illumina's P5 and P7 sequences are actually close to the annealing sites for the index primers. I suspect your P5 and P7 sequences and nearby sequences are off, hence the failure for the indexing primer to anneal and subsequently you get dark spots on the cluster (a.k.a. GGGGGG)

Hope this helps! Good luck and may the bases be with you!

[kentawan]

This upgrade really is a pain in the backside. I have no idea why Illumina couldn't make this back compatible and then use the RFID chips to determine whether to process the run as required.

I wonder what happens if someone comes wanting to compare to an older data set once we've switched to v2. They'll need to pay to re-run those old libraries on v2.

I agreed. We wrote a feedback email to their engineering team too. As you might have expected, it is a disappointing email. Lesson learnt, don't buy too many kits in one shot.

[GenoMax]

I have no idea what might cause it...... consistently.

Possibility from post #76.

[rogerzzw]

@Brian Bushnell

Illumina said it is very normal by looking the flowcell and other metrics.

[rogerzzw]

I have no idea what might cause it...... consistently.

[Brian Bushnell]

Have you looked at the images to see if there is some flowcell location correlation to the low-quality reads?

[GenoMax]

You could have a library that is partially bad?

[rogerzzw]

@GenoMax

if it is library issue, 60-70% are demultiplexed very well, I cannot explain that.

[rogerzzw]

That is not good news. That would mean there is something wrong with your bisulfite libraries since low nucleotide diversity should not have been an issue with 4 additional samples (these additional 4 samples had diverse barcodes?).

Yes, those 4 Chip-seq samples barcodes are well selected. But, the problem is I am not sure if it is due to low diversity since it is highly diverse when we combine chip-seq samples. It was still problematic when demultiplex, pretty much similar to the trial when we only have 2 samples for bisulfite sequencing.

[GenoMax]

There are a high portion (35%) reads are reading as "GGGGGG", I think it is due to reading failure, but we do not know how to pin it down, we pooled two samples each time. Unfortunately, there was one time, we pooled with other 4 Chip-seq samples and still the same.

Roger

That is not good news. That would mean there is something wrong with your bisulfite libraries since low nucleotide diversity should not have been an issue with 4 additional samples (these additional 4 samples had diverse barcodes?).

[rogerzzw]

V2 is superb! One of our run uses the same library, once on the V1 and another on V2. The V1 run's cluster density was about 170(sorry i forgotten the exact unit, while V2 was 198.

Our V1 run gave us q30 passing of about 78% while the V2 gave us 92%! Despite the higher density cluster!

Hi, Kentawan.

I am a beginner of NGS on Nextseq 500. I am primarily doing whole genome bisulfite sequencing with Nextseq 500. I saw your post on this forum and would like to learn from you.

1. based on what you said, you use V2 kit and got a very good cluster density. we used V2 but never get that high cluster density. Ours is about 140K. Can you share your experience on that? Appreciate!

2. We always had a big issue on Barcode reading, other metrics look very good. There are a high portion (35%) reads are reading as "GGGGGG", I think it is due to reading failure, but we do not know how to pin it down, we pooled two samples each time. Unfortunately, there was one time, we pooled with other 4 Chip-seq samples and still the same. Do you some idea how to fix it?

Thank you very much

Roger

[kentawan]

Were your runs using V1 or V2 reagents? There is significant differences between chemistry and run procedures between those versions. Hopefully someone with direct NextSeq experience will be able to chime in soon.

V2 is superb! One of our run uses the same library, once on the V1 and another on V2. The V1 run's cluster density was about 170(sorry i forgotten the exact unit, while V2 was 198.

Our V1 run gave us q30 passing of about 78% while the V2 gave us 92%! Despite the higher density cluster!

[rogerzzw]

Were your runs using V1 or V2 reagents? There is significant differences between chemistry and run procedures between those versions. Hopefully someone with direct NextSeq experience will be able to chime in soon.

We always use V2.

Our machine is brand new, and it had a good 1st run only on Phix. Then I started to see the tag reading problem for 4 times.

Once comparing with other people's method of library prep, we found no difference between two. But we always is quite far from expectation.

[GenoMax]

Were your runs using V1 or V2 reagents? There is significant differences between chemistry and run procedures between those versions. Hopefully someone with direct NextSeq experience will be able to chime in soon.

[rogerzzw]

@rogerzzw: I assume you have been in touch with Illumina tech support about this? What was their take?

In general tag read issues with other illumina sequencers (don't have direct NextSeq experience) happen when a sample is (borderline) overloaded. Has that possibility been eliminated?

Hi, GenoMax

I talked to tech support. We loaded as suggested (1.8pM - 2pM), but our cluster density is always lower than expect( ~140k/mm2 vs ~200k/mm2).

I learnt from several posts that the tag reading may be due to overloaded. But in our case, we actually turned out lower density, which makes me very confused.

What do you think?

BTW, illumina said it is due to low diversity libraries. But I knew there are few groups only spike in 5% Phix and got quite good results.

Appreciate!