Seqanswers Leaderboard Ad

Collapse

Announcement

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

  • Problems with recent Illumina multiplexing

    In October 2009, I prepared a paired-end multiplexed genomic DNA library using Illumina's kits and enriched for my regions of interest using Agilent's SureSelect Target Enrichment System. I cloned and Sanger sequenced a fraction of my library before proceeding to cluster generation and sequencing.

    The good news was that the Target Enrichment worked well (almost 90% in my regions of interest which was great considering that Agilent had not developed blockers specific to the multiplex adapters and primers yet, only the paired-end ones).

    The bad news was that of the 99 clones I sequenced (10 of each of 10 samples except one didn't sequence properly), only 54 had the correct adapter/primer sequences on both ends. 25 had a truncation on one end, 5 had a truncation on one end in addition to a mismatch, 1 had a truncation on both ends, 11 had one mismatch, and 3 had two mismatches. (The truncations were not all small either - only 19 of the truncations were 3 bp and under, 5 of them were 9 bp, and 1 was 10 bp.)

    In contrast, in my sequences of interest, I detected a total of 21 known SNPs (in dbSNP) and only 3 unknown SNPs. Thus, I would guess that the mismatches in the Illumina primers/adapters were not due to polymerase error, at least not the majority of them.

    I talked to Illumina customer support about this, and since nobody else had reported this problem they thought that maybe it was normal and that most people just don't clone a fraction of their library to check. They told me to go ahead and sequence the library.

    So I did sequence it using a paired-end flow cell (actually it is still running and will be done tomorrow). I ran my samples in 3 lanes and the Illumina multiplex control in a 4th lane (the other 4 lanes were for other users who had regular paired-end libraries that went through the SureSelect protocol without multiplexing.) Unfortunately, I only got 3500 clusters while the control and other users got over 100,000. I am pretty sure I didn't make any dilution errors or anything.

    I was just wondering if anyone else has (1) cloned their library and Sanger sequenced so that they can report whether they had any mismatches/truncations (2) done Illumina multiplex sequencing recently (perhaps with the same batch of adapters/primers I used) so that I can know whether I am the only one with problems with this (3) any other advice that could be useful

    Thanks!

  • #2
    Sounds like a method error. We have cloned and sanger sequenced samples that are currently on a paired-end run. The sanger QC worked well, we sequenced the prepped library before hybridization as well as the post-capture targeted and whole exome.

    For all capture protocols the sizeselect gel after ligation is important to separate self-ligated adapters that waste sequencing capacity if unremoved. To multiplex, do not introduce the index primer until the post-capture amp. We have designed our own multiplex blockers (forward and reverse of pcr primer) and will use them in the hybridization.

    Comment


    • #3
      Thank you for your reply. I did indeed do size-selection and there were not any unligated adapters that were cloned, so I am guessing that wasn't the problem. The hybridization/capture step worked well given the specificity of the clones, so I don't think that was the problem either. If you do indeed end up doing an Illumina multiplexed library, I would be curious to hear how the cloning and Sanger sequencing goes - it sounded like the one you mentioned was a paired-end one without multiplexing?

      Comment


      • #4
        Have you figured out what went wrong yet? I just read your post and had some thoughts.

        Originally posted by elizzybethy View Post
        In October 2009, I prepared a paired-end multiplexed genomic DNA library using Illumina's kits and enriched for my regions of interest using Agilent's SureSelect Target Enrichment System. I cloned and Sanger sequenced a fraction of my library before proceeding to cluster generation and sequencing.

        The bad news was that of the 99 clones I sequenced (10 of each of 10 samples except one didn't sequence properly), only 54 had the correct adapter/primer sequences on both ends. 25 had a truncation on one end, 5 had a truncation on one end in addition to a mismatch, 1 had a truncation on both ends, 11 had one mismatch, and 3 had two mismatches. (The truncations were not all small either - only 19 of the truncations were 3 bp and under, 5 of them were 9 bp, and 1 was 10 bp.)

        In contrast, in my sequences of interest, I detected a total of 21 known SNPs (in dbSNP) and only 3 unknown SNPs. Thus, I would guess that the mismatches in the Illumina primers/adapters were not due to polymerase error, at least not the majority of them.

        I talked to Illumina customer support about this, and since nobody else had reported this problem they thought that maybe it was normal and that most people just don't clone a fraction of their library to check. They told me to go ahead and sequence the library.

        So I did sequence it using a paired-end flow cell (actually it is still running and will be done tomorrow). I ran my samples in 3 lanes and the Illumina multiplex control in a 4th lane (the other 4 lanes were for other users who had regular paired-end libraries that went through the SureSelect protocol without multiplexing.) Unfortunately, I only got 3500 clusters while the control and other users got over 100,000. I am pretty sure I didn't make any dilution errors or anything.

        I was just wondering if anyone else has (1) cloned their library and Sanger sequenced so that they can report whether they had any mismatches/truncations (2) done Illumina multiplex sequencing recently (perhaps with the same batch of adapters/primers I used) so that I can know whether I am the only one with problems with this (3) any other advice that could be useful

        Thanks!
        hrm...what side of the adaptors were truncated? were the truncations always from the 5' end or 3' end? the adaptors should be protected from digestion from the 3' end if you got them from Illumina or ordered them yourself using the Sanger Centre's solution of putting a a phosphorothioate between the 2 bases at the 3' end of the adaptors. if you're missing the 5' side of the adaptors, then either they weren't properly purified after synthesis or something in your sample chewed back the 5' end. E. coli Pol I will do that (which is why we use a Klenow fragment that lacks the 5'->3' exonuclease activity).

        When I cloned out 90 transformants of my SE non-multiplexed libraries I saw a frequency of a truncated adaptor (size selection at 200-500 bp) of about 1%. these libraries at 8 pM loading concentration yielded >200,000 clusters/tile.

        if you loaded your control lane onto lane 8 you at best usually get half the cluster density as the other lanes. that could explain part of the problem. also the initial cluster density is much lower than the actual cluster density so i'd wait until the end of your run to troubleshoot it. there's a lot of information in the report files and you should try to get as many of those as possible. when it comes to dealing with companies you really need to show them that you've ruled out yourself as the major source of error (e.g. repeating it.)

        Originally posted by elizzybethy View Post
        Thank you for your reply. I did indeed do size-selection and there were not any unligated adapters that were cloned, so I am guessing that wasn't the problem. The hybridization/capture step worked well given the specificity of the clones, so I don't think that was the problem either. If you do indeed end up doing an Illumina multiplexed library, I would be curious to hear how the cloning and Sanger sequencing goes - it sounded like the one you mentioned was a paired-end one without multiplexing?
        Last edited by der_eiskern; 02-10-2010, 09:37 PM.

        Comment


        • #5
          Originally posted by upenn_ngs View Post
          To multiplex, do not introduce the index primer until the post-capture amp. We have designed our own multiplex blockers (forward and reverse of pcr primer) and will use them in the hybridization.
          Why can't you used index tags before capture? And what do you mean by
          multiplex blockers? Why and what is involved in multiplex blocking?

          I saw Agilent present some "blocking" multiplex method - but it was not clear what the technique does.

          Comment


          • #6
            The index tags (or multiplex oligos) are unique and therefore it would be necessary to design blocking oligos for each index hybridization. If you do not introduce the index until after hybridization, the same blocking oligos can be used for all reactions. The blocking oligos (forward and reverse of pcr primers) are used in hybridization to prevent daisy-chaining of genomic fragments, which will increase the percent off target capture. For post-hybrid amp use only index primer to ensure target fragments are tagged.

            Comment


            • #7
              Thanks for your explanation, but I'm a bit thick-headed. I don't quite understand in the context of target enrichment, what is blocking and what benefits there are to it.

              For example, we have been preparing genomic samples for multiplexed enrichment on a single sureselect kit. Note, I may have it wrong since I'm not doing the hands on work ourselves:

              We fragment the genomic DNA from cell samples 1, 2, 3, and 4. Then we ligate our "sequencing primers + barcode" (ie primer-barcode) to the fragments after end-repair and A-tailing. Then we evenly mix the barcoded samples together, pass them through the sureselect procedure - where we then collect the fragments, do a few cycles of PCR and ligate the final adapters.

              We sequence and we see nice read spread (~25% read per cell line). For 500 genes we get ~ 30-40X per bp read depths, and 97% target region coverage (at least one read). Agilent's claims that "hybridization bias is a problem for balanced read distributions" but that doesn't seem to come out in our numbers - perhaps because we are not doing 12X-16X multiplex!

              Can you tell me briefly the protocol involving blocking in a little more detail?
              I want to understand what the advantage is.

              Comment


              • #8
                The sureselect protocol includes blocking oligos during hybridization. For a paired-end library prep, novel sequence is introduced to the ends of genomic fragments by ligation and asymmetric pcr. The blocking oligos are used during hybridization to prevent duplex formation between the adapter regions of the genomic fragments. If the prepped library was not blocked, targeted fragments would also capture nonspecific fragments annealed to the adapters, making the targeted region a lesser proportion of the capture. As a result, a greater amount of sequencing capacity would be needed to achieve the desired depth of coverage.

                That said, the paired-end and the multiplex oligos are very similar. It is possible that you used the paired-end blocking oligos on the multiplex prepped libraries. I am interested in the method you used, please describe the pooling before hybridization and the post-hybrid amp. When considering this approach we were concerned about bias. We have seen good results with scaled-down (1/5 volume) hybridizations.

                Comment


                • #9
                  Originally posted by upenn_ngs View Post
                  The sureselect protocol includes blocking oligos during hybridization. For a paired-end library prep, novel sequence is introduced to the ends of genomic fragments by ligation and asymmetric pcr. The blocking oligos are used during hybridization to prevent duplex formation between the adapter regions of the genomic fragments. If the prepped library was not blocked, targeted fragments would also capture nonspecific fragments annealed to the adapters, making the targeted region a lesser proportion of the capture. As a result, a greater amount of sequencing capacity would be needed to achieve the desired depth of coverage.

                  That said, the paired-end and the multiplex oligos are very similar. It is possible that you used the paired-end blocking oligos on the multiplex prepped libraries. I am interested in the method you used, please describe the pooling before hybridization and the post-hybrid amp. When considering this approach we were concerned about bias. We have seen good results with scaled-down (1/5 volume) hybridizations.
                  Thanks for explaining the purpose of blocking - Now I understand the concern for duplexes formed between the adapters and other sequences.

                  I will have to ask the wet lab person for more details. I will ask if he is using paired end blocking oligos during the hybridization. My understanding is he used the standard Agilent protocol after barcoding and pooling.

                  We have written up a draft short paper on this "barcoding and pooling before enrichment" approach. It goes contrary to what Agilent presented recently at the AGBT conference , where they argue barcoding and pooling after enrichment is better.

                  Comment


                  • #10
                    Can I know how long did you hybridize your samples? I did for 24 hrs and did not get anything, although I had used the recommended amount of the prepped library (3.4 μL of 147 ng/μL). Has this happened to anyone? Any suggestions?

                    Comment


                    • #11
                      Hello everyone, glad to see this thread. I'm also going to try SureSelect together with Solexa multiplex, first pooling with barcodes and then do the hybridization (sure, we want to save money). I've also inquired both Agilent and Illumina, sadly, both of them repelled my notion of doing capturing for multiplexing. Anyway, I'll try it.

                      Question 1:
                      Actually I don't understand in the Illumina Sequence Information for multiplexing, why they use twice PCR for adding barcodes? Isn't it easier to have a longer "Multiplexing Adapter 2" and do the PCR directly from the Indices?

                      Question 2:
                      For the SureSelect, besides the blocking problem (anyway the differencing barcode is only 6 bp long, and I guess that two parts of universal blocks would also work?), does anyone have problem of doing this? What are the sequences for the Agilent Block #1, #2 and PE #3?

                      Comment


                      • #12
                        I am interested in finding out what the components of Block #1, Block #2, and Block #3 are. Does anyone know? I am sure that one of them is probably something equivalent to Cot1 DNA and the others are probably blocking oligonucleotides.

                        I have two reasons for needing to know:

                        1. I, too, am interested in barcoding BEFORE capture. There is a huge price advantage to capturing multiple samples in one capture reaction. I can see why Agilent would discourage this. This is especially true for those of us capturing only a few hundred target exons (and thus are not using much capture capacity).

                        2. I have another protocol that is using the SureSelect reagents but might need special blocking. I could just add my additional blocking to the mix, since Agilent's blocking shouldn't hurt anything, but it seems silly to add reagents that aren't needed.

                        Any suggestions?

                        Comment


                        • #13
                          I would be suspicious of the gel purification. Exposing nucleic acids to UV can nick and mutate it. I try to avoid gel purification if at all possible.

                          Has anyone substituted a size selection column such as a QIAQuick? This should remove short ligation artifacts.

                          Comment


                          • #14
                            NextGenSeq,
                            I mostly agree about gel purification, but sometimes gels are best. AMPure beads are another good way to eliminate things below 150 bp in size and have been discussed elsewhere in this forum. The QIAquick columns are good at eliminating primers, but not at eliminating that pesky 130 bp band that some here have to deal with.

                            Comment


                            • #15
                              Hi all,

                              Another question:
                              I heard that the Solexa PE and Multiplexing use different sequences for the Adaptor 2, Primer 2 and so on. And Agilent SureSelect is optimized with Solexa PE rather than Solexa Multiplexing.

                              So, how about to modify the Solexa PE adaptor/primers rather than using Solexa Multiplexing, that is, to add own barcodes in the inner side of adaptors (barcodes directly ajacent to the target sequence) and use the PE program for Solexa sequencing? This may avoid the blocking problem (only the short barcode is not blocked). Whether it will cause other problems?
                              Last edited by polyhedron; 05-03-2010, 10:10 PM.

                              Comment

                              Latest Articles

                              Collapse

                              • 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
                              • seqadmin
                                The Impact of AI in Genomic Medicine
                                by seqadmin



                                Artificial intelligence (AI) has evolved from a futuristic vision to a mainstream technology, highlighted by the introduction of tools like OpenAI's ChatGPT and Google's Gemini. In recent years, AI has become increasingly integrated into the field of genomics. This integration has enabled new scientific discoveries while simultaneously raising important ethical questions1. Interviews with two researchers at the center of this intersection provide insightful perspectives into...
                                02-26-2024, 02:07 PM

                              ad_right_rmr

                              Collapse

                              News

                              Collapse

                              Topics Statistics Last Post
                              Started by seqadmin, 03-14-2024, 06:13 AM
                              0 responses
                              34 views
                              0 likes
                              Last Post seqadmin  
                              Started by seqadmin, 03-08-2024, 08:03 AM
                              0 responses
                              72 views
                              0 likes
                              Last Post seqadmin  
                              Started by seqadmin, 03-07-2024, 08:13 AM
                              0 responses
                              82 views
                              0 likes
                              Last Post seqadmin  
                              Started by seqadmin, 03-06-2024, 09:51 AM
                              0 responses
                              68 views
                              0 likes
                              Last Post seqadmin  
                              Working...
                              X