Demultiplexing reads with bad i7 indices

Question

[Garrett]

I'm trying to demultiplex RADseq reads using bcl2fastq, but I'm having a lot of trouble with the i7 indices. I tried a bunch of combinations for i5 and i7 using complement, reverse complement, etc., but the reads couldn't demultiplex. I ran --create-fastq-for-index-reads and found that all the i7 indices appear to be all G's, AAATAAAA, or something else with a lot of repeats.

I was thinking the i7 indices might be in a different position, but I'm not sure. Another problem is that bcl2fastq isn't adding the i5 indices to the header, so I can't try separating by i5 and then going from there. Any advice?

Answer 1

[GenoMax]

I ran --create-fastq-for-index-reads and found that all the i7 indices appear to be all G's, AAATAAAA, or something else with a lot of repeats.

If the i7 index failed on sequencer then nothing can be done to recover this run. Your sequencing provider should have re-run this pool.

Another problem is that bcl2fastq isn't adding the i5 indices to the header, so I can't try separating by i5 and then going from there

This could be done by editing RunInfo.xml file to say that i7 read cycle is not an index or by using a --use-bases-mask directive. But if you can't separate the i7 indexes then this is not going to help.