Strand bias in amplicon sequencing

[Ingeneious]

I am seeing random strand bias in amplicon reads on an Illumina MiSeq run. Some of the amplicons have effectively complete dropout of the forward or reverse read. This occurs across every sample on the run.

I have used this library prep kit before and did not see these results before. Also, the bias occurs in regions that I would not expect this to happen (50% GC).

Has anyone seen this before or have any ideas what could be causing this? Is there anything in the library prep process that might cause this?

[nucacidhunter]

Is there anything in the library prep process that might cause this?

Would you give more info on library prep method, indexing, fragment size and number of sequencing cycles.

[Ingeneious]

TruSeq DNA HT, 2 x 8bp, 200bp, 2 x 150 PE.

[nucacidhunter]

I thought it was a PCR with target specific primers, but from your description I assume a shotgun DNA library. It is not clear to me what "Some of the amplicons have effectively complete dropout of the forward or reverse read" and how you have come to this conclusion. I wonder if it means that you did not get a read2 for some read1s.

[Ingeneious]

This is ligation-based amplicon sequencing, so no shotgun involved.

What could cause read2 dropout? %Q30 scores were all >90%.

[nucacidhunter]

Once a cluster passes filter in R1, indexing read(s) and R2 will collect data for that cluster even if it physically is not present. There should be still sequence data for those low quality R2s and it should not be a blank sequence file.
Possible reasons:
1- Overclustering or close to the max density specially with large amplicons
2- Low diversity regions in R2
3- Issues with sequencing primers or reagents
4- Issues with instrument hardware

Time to call Illumina tech support.

[Ingeneious]

Thanks for your input, nucacidhunter.