Seqanswers Leaderboard Ad

Collapse

Announcement

Collapse
No announcement yet.
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • HiSeq 2500 Rapid run dual index problems

    Our lab is having problems getting successful dual index runs on the HiSeq 2500 rapid module (upgraded from the HiSeq 2000). Single index runs work, but on dual index runs the quality craps out before it can read the second index.

    Illumina told us that it could be a template diversity issue, but spiking in PhiX up to 30% did not resolve the issue.

    I am wondering if this is specific to our HiSeq, or if other labs have noticed similar problems.

    Thanks,
    Brett

  • #2
    Brett: What cluster density are you running at?

    If your flowcells are (on the border or) over-clustered then generally the sequences look fine but the tag reads suffer badly. They are rather sensitive to cluster density.

    Comment


    • #3
      We are getting 600-700K clusters. The first read and index look fine, but the second index and reads fail (all Ns).

      Comment


      • #4
        When we have had issues the reads have always been ok (only the tag reads have problems so effectively you can't demultiplex the data).

        I don't think we have had any general problems otherwise with a couple of HiSeq 2500 (which were upgraded from 2000).

        BTW: Your cluster density looks ok.

        Comment


        • #5
          Is there is a chance it could be sample related. Assuming these are either HT or Nextera based, are they different preps (another company? someone else doing the preps?) than say your single indexed libraries, which I assume are Truseq LT?
          Are the runs that have dual indexing, also paired-end runs? If they are paired-end then I think the quality may depend on length of the libraries because the i5 index primer is grafted in comparison to dual-index single read. So it needs to be able to bridge (pg13 on user guide B). If PE, then how does the quality of the 4th read look?
          I concur that the density seems fine.
          I'm sure you are asking for future runs, but FYI, there is a chance you could save the data from previous runs. Are any of the i7 indices the same? You may be able to reconfigure your sample sheet and/or BCLtoFastq to salvage the samples that are unique in the i7 reads.

          Comment


          • #6
            This is almost certainly a library or adapter problem. More details required about library prep and adapter configuration.

            Comment


            • #7
              We have been prepping these dual index paired end libraries on our IntegenX Apollo 324, which essentially uses a TruSeq HT-like chemistry. I believe the adapters were purchased through IDT.

              These libraries give us good sequence on the MiSeq, so I don't think it is a library prep issue.

              Comment


              • #8
                Originally posted by bbeitzel View Post
                These libraries give us good sequence on the MiSeq, so I don't think it is a library prep issue.
                Keep that option open.

                We have had some instances where libraries failing on a HiSeq have worked on a MiSeq run, when checked after the fact.
                Last edited by GenoMax; 04-24-2014, 08:13 AM.

                Comment


                • #9
                  Hi, I am having this exact same problem right now with 4 libraries. 2 ChIP, 2 RNA. Good clustering and quality scores, and then drop down to zero with second read. I have sanger-sequenced these libraries and the clones I picked have the correct structure and indexing.
                  Sequencing core is at a loss.

                  Comment


                  • #10
                    We have had a couple of runs on rapid flowcells on the 2500, PE100, with similar results. Great read 1 and paired end read, but awful index reads. Only got basecalls on something like 8M out of 120M total. Similar libraries have been sequenced on the high output flowcells without issue. Has anyone figured out this issue when it has happened to them?

                    Comment


                    • #11
                      Originally posted by Aaron Cooper View Post
                      We have had a couple of runs on rapid flowcells on the 2500, PE100, with similar results. Great read 1 and paired end read, but awful index reads. Only got basecalls on something like 8M out of 120M total. Similar libraries have been sequenced on the high output flowcells without issue. Has anyone figured out this issue when it has happened to them?
                      @Aaron: Anytime something like this happens get in touch with Illumina tech support right away. They can remotely take a look at the run and diagnose to see if it is an instrument/reagent issue. In both cases you will get replacement reagents.

                      That said "similar libraries" have sequenced well does not mean the ones you have on hand are of equally good quality. If the same library has sequenced well before then that is a different case.

                      Comment


                      • #12
                        Okay, there is the perennial "index diversity/color balance" issue that seems to have re-emerged with a recent HCS (HiSeq Control Software) version. In earlier days (when the HiSeq first was released) this was a serious issue that Illumina mentioned nowhere.

                        Basically if you didn't have nucleotide (color) balance among your indexes, there would be trouble with demultiplexing. This was a problem at multiple levels, but one was that if the instrument saw no signal when it shined its, for example, green laser on a lane because the all the indexes in that lane just happened to have encoded red-laser-bases at that position, then the machine flakes out. Back then (years ago) this was a really bad flake-out -- like the instrument changed its focal plane trying to pick up signal and ended up way, way out of focus.

                        Around the time this was fixed, a couple of years back, Illumina also began warning people of the issue in their literature. Meaning everyone started stripping their gears attempting to achieve perfect index color balance, while the software no longer had a problem with lack of balance.

                        But other than the irritation of seeing that people were wasting time on an issue that no longer was an issue, these were halcyon days.

                        Then a recent version of HCS (ironically, I think it was the one that allowed the HiSeq to deal with high-bias libraries like the MiSeq was able to) seems to have resulted in at least a partial return to the bad old days. Nowhere near as bad as it once was, but might be causing issues again. We've mainly seen it in cases where at least the first 2 bases of all indexes are the same in a lane.

                        To avoid this read, for example, the relevant section of the Nextera library construction manual on how to achieve color balance.

                        --
                        Phillip

                        Comment


                        • #13
                          Thanks for the ideas, GenoMax and Phillip.

                          We had 57 indexes in this run, and they were well-balanced. The only idea I have is that HCS doesn't like high PhiX spike-in, because it leads to lots of dark clusters during index cycles. Our libraries have very low diversity at the beginning of read 1, so we use ~50% PhiX. In this run, it actually ended up being closer to 60%. Illumina is replacing the run, but we still haven't gotten any good ideas from them about the problem. I'm thinking about running this library again with a different spike-in that has indexes.

                          Comment


                          • #14
                            Actually, with newer versions of HCS, you don't get any advantage from adding phiX beyond 10% or so.
                            Maybe something screwy is going on with your library titration and you actually added >90% phiX?

                            --
                            Phillip

                            Comment


                            • #15
                              When I say very low diversity, I mean the first 20 nt are identical. Maximally low diversity! The core here indicated that Illumina was still recommending 50% for this situation-- is that indeed no longer true?

                              The 50-60% figure I gave was based on looking at the run data, so it should be accurate.

                              Comment

                              Latest Articles

                              Collapse

                              • seqadmin
                                Strategies for Sequencing Challenging Samples
                                by seqadmin


                                Despite advancements in sequencing platforms and related sample preparation technologies, certain sample types continue to present significant challenges that can compromise sequencing results. Pedro Echave, Senior Manager of the Global Business Segment at Revvity, explained that the success of a sequencing experiment ultimately depends on the amount and integrity of the nucleic acid template (RNA or DNA) obtained from a sample. “The better the quality of the nucleic acid isolated...
                                03-22-2024, 06:39 AM
                              • seqadmin
                                Techniques and Challenges in Conservation Genomics
                                by seqadmin



                                The field of conservation genomics centers on applying genomics technologies in support of conservation efforts and the preservation of biodiversity. This article features interviews with two researchers who showcase their innovative work and highlight the current state and future of conservation genomics.

                                Avian Conservation
                                Matthew DeSaix, a recent doctoral graduate from Kristen Ruegg’s lab at The University of Colorado, shared that most of his research...
                                03-08-2024, 10:41 AM

                              ad_right_rmr

                              Collapse

                              News

                              Collapse

                              Topics Statistics Last Post
                              Started by seqadmin, Yesterday, 06:37 PM
                              0 responses
                              11 views
                              0 likes
                              Last Post seqadmin  
                              Started by seqadmin, Yesterday, 06:07 PM
                              0 responses
                              10 views
                              0 likes
                              Last Post seqadmin  
                              Started by seqadmin, 03-22-2024, 10:03 AM
                              0 responses
                              51 views
                              0 likes
                              Last Post seqadmin  
                              Started by seqadmin, 03-21-2024, 07:32 AM
                              0 responses
                              68 views
                              0 likes
                              Last Post seqadmin  
                              Working...
                              X