Announcement

Collapse

Welcome to the New Seqanswers!

Welcome to the new Seqanswers! We'd love your feedback, please post any you have to this topic: New Seqanswers Feedback.
See more
See less

MiSeq Cluster Density Revisited

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • MiSeq Cluster Density Revisited

    This is discussed ad nauseum, but I can't seem to find the answer I'm looking for so here goes! Over the last two months my cluster densities have gone from 800-1000 K/mm2 to 500-600 K/mm2. Same v2 chemistry, same library prep (TruSeq-like), same KAPA quantification. I've been clustering with 8 pM libraries and tried going up to 8.5 pM but it hasn't made a difference.

    What concentration are most of you clustering at? I'd like to get back up to the ~900 K/mm2. And can NaOH go bad? I'm thinking that maybe it's not denaturing properly. Is it possible to buy it in small aliquots so that I can open a new tube/container every couple of runs?

  • #2
    Our clustering is a bit variable, but we're generally in the 700-1000K range for all runs, with the average over our last few being ~850K. This is a bit lower than what the MiSeq can obviously support, but we'd rather under cluster a little than over cluster and get terrible data back.

    As for the NaOH, according to Illumina it can "go bad". I've been storing 25ul aliquots of 2N NaOH in our freezer and each time we go to denature libraries we pull out a new aliquot and dilute it to 0.1N instead of 0.2N. We used to have a ton of problems getting consistent clustering, but since we've switched to this method we've been more consistent even though we're still working with lots of different library types that require different final pM concentrations in order to get appropriate clustering.

    I'll note that we still don't use the Kappa kit for quantification and only do Qubit HS DNA and Bioanalyzer, so that obviously introduces some variability in our clustering since we're not quantifying just fragments that would bind to the flow-cell.

    Comment


    • #3
      Regarding NaOH, I wonder what the best way is to store it. Does anyone store it under mineral oil at room temperature? Or is freezing an effective method to prevent interaction with CO2 in the air?

      Also, mcnelson.phd, have you had any conversations with Illumina about your better/more consistent clustering with reduced [NaOH]?

      Comment


      • #4
        Originally posted by clintp View Post
        Also, mcnelson.phd, have you had any conversations with Illumina about your better/more consistent clustering with reduced [NaOH]?
        The recommendation to use 0.1N NaOH as the starting point for denaturing actually came from our FAS when we were discussing our problems of inconsistent clustering. At first we thought we were having fluidics problems because some runs had densities in the 100-200K range, but that was ruled out as a problem. Our FAS told us that the 0.2N value is really overkill, and that they were recommending the 0.1N starting value to their other sites.

        Since we've started using 0.1N for everything, and using the frozen 25ul aliquots I mentioned above, we've had very consistent clustering for runs that only feature a single library type (e.g. only 16S amplicons or only Nextera genomes). Most of our variation comes up when we mix different library types and try to compensate for the fact that smaller fragments cluster better by diluting libraries to different final pM concentrations before pooling them to load into the cartridge.

        Last words I really have are that we have also given up entirely on using the XT library normalization and denaturing process and instead take the post-PCR cleaned libraries and treat them like standard Nextera libraries. Since we've started doing that we haven't had the horribly uneven pooling or over clustering issues we had when using the official normalization and denaturing protocol.

        Comment


        • #5
          We have been freezing 1 ml aliquots of 1N NaOH (no mineral oil). We toss the aliquot after the 2nd thaw. Dilute to 0.2 N according to Illumina's protocol. Never see any problems with clustering doing it this way. Good luck.

          Comment


          • #6
            Originally posted by clintp View Post
            Regarding NaOH, I wonder what the best way is to store it. Does anyone store it under mineral oil at room temperature? Or is freezing an effective method to prevent interaction with CO2 in the air?

            Also, mcnelson.phd, have you had any conversations with Illumina about your better/more consistent clustering with reduced [NaOH]?
            I stored original Illumina's aliquiot of 2N NaOH just on the bench, tightly closed, making single use 20 ul dilutions every time. Did not notice much difference after a year or so. Indeed, how much CO2 is in the air over solution in the capped tube even if it is opened and recapped? Hard to imagine something less complicated than NaOH, may be salt or water...

            Comment


            • #7
              I made some 2N NaOH when our miseq arrived a year ago and it has lived on bench since then in a 1.5mL snap top tube. I dilute fresh 0.2N from this stock each time. No trouble!!

              Also, last time I denatured phix was May. I just keep the same 20pM stock in the freezer. Probably 25 runs or so since then, no problems with PhiX either.

              And to the original question about clustering, we used to load at lower concentrations (~8-12pM depending on application), but around June our cluster density started to drop off. Illumina came out and cleaned our lines and upgraded the valve and back in business. We now load between 14-18pM and typically see clustering between 900-1200k. Every now and then we botch it a little, usually when a user supplies a library with adapters remaining. For amplicon runs, we load a bit lower since these seem to cluster more densely. My picogreen and qPCR quants are usually in good agreement. I wouldn't trust a bioanalyzer quant on a library that has been amplified at all (unreliable due to pipet error - just 1ul, and the nonvisualizable heteroduplex component). Generally increasing load concentrations by just a hair (0.5pM) has little to no effect, probably due to variability in library quantification. If you are clustering at 500k with 8pM, I would bump up to 12pM for the next run and see where that puts you. It's not exactly a linear relationship between loading concentration and cluster density, but not far off either. You might want to go as high as 16pM. For Kozarewa and Turner preps (truseq like), we load at 18pM.

              For qPCR we don't use the Kapa kit. Great company (check out the 2G Fast polymerase), but how hard is it to make qPCR standard, especially when you are making Illumina-compatible amplicons? Just take a bit of a previous amplicon library (length should be known - I use 16S v4 construct), and reamplify it to produce more. Bead clean it (see this post: http://enggen-nau.blogspot.com/2013/...-cleanups.html), pico-quantify, and std made. Use P5/P7 sequences as primers. Can also use these for QC checks during other preps (TSCA, Nextera, etc).

              Comment


              • #8
                One more thing regarding NaOH. It is ridiculously easy to make. 400g NaOH pellets, dissolve in H2O (exothermic), bring to 1L (10N solutioN). Cap tightly and it is good for years.

                Comment


                • #9
                  Dear All,
                  i'm running TruSight Cancer Nextera Enrichment protocol for libraries (genomic DNA) on MiSeq and end up in quite low (although very good quality reads) cluster density 180K with 0,2N NaOH. The length of libraries on average was 1000 bp which is quite long i think, what could impede the generation of clusters on flowcell. The input of libraries I made according the illumina protocol suggestion ~20nM. What would be the suggestions to improve cluster density? Should i use 0,1N NaOH (this is actually the quantity which is in the arrays preparation protocol)?

                  Comment


                  • #10
                    In regard to eliminating effect of remaining NaOH I found a very simple solution. Make solution of 1M acetic acid in water, it perfectly stable on a bench. Add 1 ul to the final dilution of your library in HT before loading into the library well - cluster density jumps several fold. I overclustered a few times before I realized I got to be more conservative with this modification.
                    BTW, I prefer not to waste library, so I use exact amount recalculated back to 600 ul of final volume from the protocol recommendations. So, for example, to get 600 ul of 50 pM I simply combine 3 ul 10 nM with 3 ul of 0.2 M NaOH, wait, then dilute straight with 594 ul of HT, add 1 ul of 1 M AcOH and load the well. One can hit density 3000 at this concentration, though.

                    Comment


                    • #11
                      Originally posted by ramujana View Post
                      Dear All,
                      i'm running TruSight Cancer Nextera Enrichment protocol for libraries (genomic DNA) on MiSeq and end up in quite low (although very good quality reads) cluster density 180K with 0,2N NaOH. The length of libraries on average was 1000 bp which is quite long i think, what could impede the generation of clusters on flowcell. The input of libraries I made according the illumina protocol suggestion ~20nM. What would be the suggestions to improve cluster density? Should i use 0,1N NaOH (this is actually the quantity which is in the arrays preparation protocol)?

                      We've got similar issue with the Miseq, low density at 122K with 0.2N NaOH according to Miseq user guide. It's a polyA mRNA lib using the NEB kit, with 750bp length on average. The final concentration of loading lib is 11pM.

                      Is it the long lib that cause the problem or the NaOH concentration?
                      any suggestion?
                      Cheers,
                      Tanya

                      Comment


                      • #12
                        when we get the illumina training one year before, the FAS have told us to use 0.1N NaOH. it is interesting that yaximik use 1M acetic acid before loading on FC, have anyone tried HCl?

                        Comment


                        • #13
                          HCl is too strong acid I think, it will skew pH significantly. Obviously, one ul of 1M acetic acid is in excess to remaining NaOH, but it is much weaker and rather create acetate buffer.

                          Comment


                          • #14
                            Yaximik, I am suddenly finding myself interested in your acetate mod. I have some libraries that just don't want to reach 4nM. Your mod suggests adequate clustering can be had with lower concentration libraries, however, I notice your final NaOH concentration is the same as the rest of us (0.001N), and yet you are still adding acetic acid. Has this improved your clustering consistency? What is the lowest concentration library you have denatured and achieved useful cluster density?

                            Comment


                            • #15
                              Addition of acetic acid likely will not, or very slightly may improve clustering above what is achieved with final NaOH at 1 mM. The mod was designed to take care of higher NaOH concentrations, which happens if you mix libraries at low concentration with equal volume of 0.2 M NaOH. The lowest concentration I used was 2.5 nM (for v3 kits). I used 3 ul of library and 3 ul of 200 mM NaOH. After 5 min, I hung 1 ul of 1 M AcOH on the tube wall, then diluted all straight with 600 ul of HT, and loaded into the well. In two cases I got ~1400 cl/mm2. Sometimes I intentionally dilute 10 nM libraries two or three times to use more reliable volumes than, say 0.75 or 1 ul, and use accordingly 2-3 times more 0.2 M NaOH. Perhaps you could use 1.5 ul of your 4 nM libraries. I do not see why one cannot use much lower library concentrations, as AcOH is added in huge excess to NaOH anyway.

                              Comment

                              Working...
                              X