What happens if a custom read 2 sequencing primer fails?

[bilyl]

I'm working on a special library prep scheme for a specialized project that uses a custom read 1 and read 2 sequencing primer. In principle, the first read will always work, but only the correctly formed fragments will have a read 2 complementary sequence. In both cases, it will still have the P5/P7 sequence for bridge/cluster amplification.

How would this look in a Miseq/Hiseq run? Would it give me a bunch of Ns in my read 2 since there would be no intensity associated with the location of the cluster (and probably kill my Q-score)? Would this affect any kind of registration issues for the other properly formed clusters? I'm guessing these types of reads would be easy to filter out post-run...

Or would it just be much easier to reverse the read on which an improperly formed fragment may form (eg. read 1 may fail sometimes)? In this case the cluster would just fail to register in RTA for the first few cycles. I've already made some preliminary libraries so it would be more work to "reverse" everything.

[nucacidhunter]

It would be easier to reply if there was more information about library structure and aim. Generally, if read2 sequanceable clusters are below a treshold, camera will not be able to focus and sequencing will abort. One way to work around this would be to spike-in 25% Phix. It also depends on which system is used.

[bilyl]

It would be easier to reply if there was more information about library structure and aim. Generally, if read2 sequanceable clusters are below a treshold, camera will not be able to focus and sequencing will abort. One way to work around this would be to spike-in 25% Phix. It also depends on which system is used.

I can't exactly say too much about what we're doing... but as a ballpark, if 10-15% of clusters "disappeared" during Read 2, would that cause the run to abort?

[GenoMax]

I can't exactly say too much about what we're doing... but as a ballpark, if 10-15% of clusters "disappeared" during Read 2, would that cause the run to abort?

As long as there is signal from the remaining 80% the run should (in theory) continue. You could try a test run using a 50 bp PE MiSeq kit (if you have access to a MiSeq). That would be the cheapest option to validate your method.

[bryanbriney]

Would it give me a bunch of Ns in my read 2 since there would be no intensity associated with the location of the cluster (and probably kill my Q-score)?

NextSeq would give you reads with all Gs.

[nucacidhunter]

I can't exactly say too much about what we're doing... but as a ballpark, if 10-15% of clusters "disappeared" during Read 2, would that cause the run to abort?

The answer is no. RTA forms template and filters them during first 25 cycle of read1. So any cluster passing filter will have a FASTQ entry for both read1 and read2. If some of read1 clusters do not have a 2nd read sequencing primer binding site, obviously there will be no signal from them. But, RTA still assigns them base calls which is result of signals from neighbouring clusters. Those will have low Q scores. This happens during sequencing fragments shorter than cycle number as well when the template, adapter and flow cell lawn runs out. The problem is that low Q scores can be for other reasons as well and you will have to sort this out to identify the real failed read2. Camera needs aroun 200k cluster density for focosing, so if your read2 clusters are above this treshold sequencing will continue and will not be aborted.