Failed NextSeq 500 v2 runs w/ custom sequencing primer

[kakseq]

Hi,

Any tips regarding the use of a custom primer on the NextSeq would be appreciated.

We are trying to run a multiplexed set of 3' RNASeq libraries with a custom sequencing primer on a NextSeq with V2 chemistry with a 75 bp high output kit. The libraries were designed so that all reads come from the 3' UTR just upstream of the poly-A tail. We ran a set of these libs successfully on the MiSeq, but it has failed three times with the NextSeq. The primer contains 18 T's to cover the 18 A's that are at the end of the first strand synthesis primer which is used to make cDNA (see below). We have tried this custom primer with and without the 3' phosphorothioate modification between the final two T's. Without the phosphorothioate we had high phasing (~0.3), low pass filter rates, high error and medium to low intensity, with a total output of 189 M reads. With the phosphorothioate between the last two T's we had very low intensity and only 10 M reads passed filter, but normal cluster density and normal phasing around 0.12. In both cases the barcode reads look excellent and the cell was not over clustered. There were multiple successful runs on this instrument between and after these failed runs.
We were worried that the long stretch of T's might be melting off, but I know the MiSeq runs at 65C while the NextSeq runs at 60 C. So, if melting were the problem then the run should have been worse on the higher temperature of the MiSeq.

The custom primer is underlined here. It largely overlaps with the Illumina read 1 primer:
5’AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGATCT-TTTTTTTTTTTTTTTTTT-Insert…
3’TTACTATGCCGCTGGTGGCTCTAGATGTGAGAAAGGGATGTGCTGCGAGAAGGCTAGA-AAAAAAAAAAAAAAAAAA-Insert…


Here are the stats from the latest failed run:
Run Summary

Level Yield Total (G) Projected Total Yield (G) Aligned (%) Error Rate (%) Intensity Cycle 1 % >= Q30
Read 1 11.2 11.2 0 0 3248 79.3
Read 2 0.7 0.7 0 0 5418 96.4
Total 11.85005 11.85005 0 0 4332.993 80.39454

Read 1

Lane Tiles Density (K/mm2) Clusters PF (%) Phas/Prephas (%) Reads (M) Reads PF (M) % >= Q30 Yield(G) Cycles Err Rated Aligned (%) Error Rate (%) Error Rate 35 cycle (%) Error Rate 75 cycle (%) Error Rate 100 cycle (%) Intensity Cycle 1
1 216 210 +/- 61 29.43 +/- 18.55 0.150 / 0.084 136.33 35.55 77.6 3 +/- 0 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 3294 +/- 641
2 216 236 +/- 37 27.98 +/- 20.33 0.089 / 0.052 152.83 40.4 81.4 3 +/- 0 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 2927 +/- 469
3 216 184 +/- 80 29.13 +/- 18.35 0.082 / 0.075 119.12 28.78 79 2 +/- 0 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 3616 +/- 943
4 216 217 +/- 68 22.66 +/- 18.45 0.172 / 0.089 140.47 28.42 78.5 2 +/- 0 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 3154 +/- 1160

Read 2

Lane Tiles Density (K/mm2) Clusters PF (%) Phas/Prephas (%) Reads (M) Reads PF (M) % >= Q30 Yield(G) Cycles Err Rated Aligned (%) Error Rate (%) Error Rate 35 cycle (%) Error Rate 75 cycle (%) Error Rate 100 cycle (%) Intensity Cycle 1
1 216 210 +/- 61 29.43 +/- 18.55 0.000 / 0.000 136.33 35.55 96.2 0 +/- 0 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 4668 +/- 2164
2 216 236 +/- 37 27.98 +/- 20.33 0.000 / 0.000 152.83 40.4 96.8 0 +/- 0 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 5927 +/- 1886
3 216 184 +/- 80 29.13 +/- 18.35 0.000 / 0.000 119.12 28.78 95.7 0 +/- 0 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 5167 +/- 2191
4 216 217 +/- 68 22.66 +/- 18.45 0.000 / 0.000 140.47 28.42 96.9 0 +/- 0 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00 0.00 +/- 0.00

[nucacidhunter]

Illumina’s recent NextSeq System Custom Primers Guide might give some ideas to improve run output:http://support.illumina.com/content/...15057456-d.pdf

[luc]

Hi kakseq,

did you figure out the problem? We are also having problems over here - way too low cluster densities with custom seq primers.

Thanks!

[kakseq]

It seems that the long stretch of TTTT's in our custom primer was the problem. We couldn't solve this on the NextSeq even after extensive chatting with Illumina and the kit manufacturer. We actually ended up switching the way our libraries are made so read 1 comes from the other end of the fragment so no custom primer is needed so we could keep using our NextSeqs.
There are some nucleic acid modifications which may help, like locked nucleic acids (LNAs) or Super-T, but we haven't tried this. After half a dozen failed runs I don't have much more appetite to tinker with this problem.

Here is IDT's smorgasboard of TM-altering modifications which may be of interest.

https://www.idtdna.com/site/Catalog/Modifications/Category/7

If you try out the modifications please let me know how this works out for you.

[luc]

Thanks a lot kakseq!

The custom seq primer that we need to work with at the moment also has two stretches of Ts (a 5T and a 4T stretch)/