Sequence transposon flanking region

[Akira]

I am new to NGS first of all. I would like to sequence transposon flanking region, thus during PCR amplification step, instead of using primers supplied by Illumina TruSeq, I will use own customized primer (tailed with Illumina adapter sequence) which will bind specifically to the transposon end and start amplification. Lots of issue I need to settle for sample prep. But now my doubt is: during sequencing run, is it possible if we use own sequencing primer (in which I plan to design primer which binds to the transposon end region)? Say if I wanted to run on HiSeq2000. If I use standard sequencing primer, most of the first ~20bp of the reads will be the same (the transposon end) and how will this affect the coordination of clusters during first few cycles?

Sorry for the poor English used.

[HESmith]

Low complexity sequence during the first four cycles will adversely affect cluster calling. Sequencing with a custom primer is fine, as long as the Tm is comparable to the standard sequencing primer.

[Akira]

Many thanks to the reply.

If paired end, will the cluster calling be repeated for read 2? Or once the first four cycles of read 1 have coordinated the clusters position, it doesn't matter if read 2 starts with low complexity sequence? For sure more than 10bp of the first read will be low complexity sequence (as mentioned the transposon end), I was thinking if I design the primer in a way the transposon end is actually located on read 2, will it still affecting the cluster coordination?

Custom sequencing primer can replace the original primer? I have been told that it comes in a master mix together with other reagents needed for sequencing run. Is it true?

[krobison]

By low complexity, it is meant low diversity of the sequence -- your primers may cause issues due to the fixed targeting region. One way to try to address this would be to use a pool of primers for the amplification step, with different lengths of spacer between the targeting region and the Illumina tail -- this will prevent the constant region from always being in register & should help with the cluster calling (idea is not mine; someone published a scheme in which the multiplex tags used variable lengths to solve the problem)

[Akira]

Many thanks for the reply. Cluster calling happens in both read 1 and read 2 for paired end sequencing or only read 1? Most of the clusters will be filtered out if the clusters calling was not good enough, am I right?

[HESmith]

Cluster calling is based solely on the first four cycles of read one, so low diversity of read two will not affect the data.

Custom primers can be used during clustering on the cBOT. Select the appropriate protocol and load the primers in tube strip. See pp. 53-4 of the cBOT user's guide for more info.

[Akira]

Thanks HESmith. In this case, I can actually revert back the concept by starting to sequence from the other side of the fragments, which is the genomic DNA as read 1, then the transposon end with low sequence complexity as read 2. Am I right? Then only I filter out those sequence reads without any transposon end tag on read 2 using bioinformatic tools, possible? Because I am only interested in those reads tagged with transposon end, it doesn't matter the tag is on read 1 or read 2. I hope I get the whole picture right.

Your advice is much appreciated.

[Janiz]

Hi Akira,

I am considering using Epicentre's (now Illuminas) Nextera DNA sample prep kit. It seems there's been some changes to this kit this year and the index primers are now sold separately. Do you possibly know the sequences of the tagged DNA after the tagmentation and cleanup? Is it possible to create our own primers to measure tagmentation efficiency without having to order the index kit? In this new protocol it seems you need the index kit (at least Index 1 and 2) to be able to perform the PCR amplification step and add on the adapters (after which you could do a QPCR and measure tagmentation efficiency) although they do provide the PCR buffer and primer mix?

Thanks!

[pmiguel]

Cluster calling is based solely on the first four cycles of read one, so low diversity of read two will not affect the data.

While the former point is correct, I think the latter point no longer is. Prior to a firmware/software update last year (I think the one that allowed use of v3 chemistry) the reverse PE read apparently just used the focal settings acquired during the first cycle of read1. This was good in that diversity was not an issue for this read. But it was bad if there was an index read before the reverse PE read and this index read had hosed the focal settings due to its low diversity. (Like if you had some indexed lanes and some non-indexed, or single index lanes.)

Under such circumstances the work-around was to exit HCS entirely prior to turn-around. Upon re-entering HCS, the software would force a focal settings calibration prior to read 2. So Illumina changed the software so it always does the re-calibration prior to read 2. But that leaves you with the low diversity bug for both read 1 and read 2.

At least this is my interpretation of where things stand currently. Corrections welcome, of course.

--
Phillip