Tweet
I produced a single digest RAD library using Sbf-I, a rare cutter, following a variation of the Etter (2011) protocol available on Michael Miller's website https://sites.google.com/site/millermichaelryan/. I am working with Mytilus spp., which has a large genome and low GC content. I got back lots of data.
However, I am running into a problem. When I use the process_radtags v. 0.9996 script available in the Stacks package, the majority of p1 reads are filtered out due to "ambiguous RADs", and then an almost equivalent number of p2 reads are "orphaned".
Total Sequences 37328778
Ambiguous Barcodes 0
Low Quality 548144
Ambiguous RAD-Tag 15059135
Orphaned Paired-ends 15529880
Retained Reads 6191619
How are all these ambiguous RAD-tags created? Non-specific ligation of adaptors?
When I proceed with analysis in Stacks with only the process_radtags retained reads (6191619 in the example above), I retrieve ~60 K stacks per species, which makes approximate sense given the number of cut sites I expected with a genome of this size and GC content (~15-20 K).
When I proceed with analysis with all high quality reads (a little less than 37328778 in the example above), I retrieve ~ 2 million stacks per species, which does not even remotely match the number of cut sites expected. I have also tried double digest RAD seq on this species with a Pippin Prep size selection (following the Peterson protocol), in which PCR failed. This gave me further reason to believe that there more like 30,000 SbfI cut sites in the Mytilus genome than 1 million cut sites.
Has anyone found anything similar: Majority of reads turning out to be ambiguous RAD-tags? How are all these ambiguous RAD-tags created?
Thanks for any help!
- Norah Saarman
However, I am running into a problem. When I use the process_radtags v. 0.9996 script available in the Stacks package, the majority of p1 reads are filtered out due to "ambiguous RADs", and then an almost equivalent number of p2 reads are "orphaned".
Total Sequences 37328778
Ambiguous Barcodes 0
Low Quality 548144
Ambiguous RAD-Tag 15059135
Orphaned Paired-ends 15529880
Retained Reads 6191619
How are all these ambiguous RAD-tags created? Non-specific ligation of adaptors?
When I proceed with analysis in Stacks with only the process_radtags retained reads (6191619 in the example above), I retrieve ~60 K stacks per species, which makes approximate sense given the number of cut sites I expected with a genome of this size and GC content (~15-20 K).
When I proceed with analysis with all high quality reads (a little less than 37328778 in the example above), I retrieve ~ 2 million stacks per species, which does not even remotely match the number of cut sites expected. I have also tried double digest RAD seq on this species with a Pippin Prep size selection (following the Peterson protocol), in which PCR failed. This gave me further reason to believe that there more like 30,000 SbfI cut sites in the Mytilus genome than 1 million cut sites.
Has anyone found anything similar: Majority of reads turning out to be ambiguous RAD-tags? How are all these ambiguous RAD-tags created?
Thanks for any help!
- Norah Saarman