The ddRAD data we are producing have an in-line barcode (within the sequence read) in 3'and an index barcode in 5'end (at the read heading in the fastq file). (http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0037135).
If I go for single-end sequencing, how can I tell apart two samples with the same in-line index in 3' but different indexes in 5'?
You will need to demultiplex each P7 index set separately. Do you currently have a separate sequence file for each P7-end index? If not you could sort them into different files with a shell command. The next step would be to sort by the inline barcode. Both of the most popular RAD processing pipelines (STACKS and pyRAD) I believe take as input a tab-delimited file arranged like so:
SampleName1 (tab) ACGTG
SampName2 .....
Alrhough you could do this part as well with your own script, but both of the above also will attempt to demultiplex reads with errors introduced into the barcode, and also do a lot of nice quality filtering for you. The end result should obviously be a separate fast file for each sample.
Personally I prefer pyRAD and use it very often so if you need any help with pyRAD feel free to PM me, and I also recommend you check out the user group as the lead developer Derek Eaton is very responsive and helpful. https://groups.google.com/forum/m/#!forum/pyrad-users
Thank you for your tips.
I still have a question. If you go for single end sequencing you still get the P7 index? or you need to do paired end to retrieve the P7 index?
Yes with single end sequencing you will still get the P7-end index sequence. Be sure to provide those P7 index sequences for each pooled sample to your facility. The raw output the sequence facility provides usually is in the form of a separate fastq file for each P7 index, which you can then demultiplex for the P5 index (the first N bases of each read).
- Badges
- Science topic
- Similar topics
- Composition Studies
- Articles