Hi Peter,
Are you asking for a tool to interleave to FASTQ or FASTA files with matching entries (with matching names in the same order) into one file which alternates forward then reverse read?
Yes, indeed, this is what I am proposing.
Would you prefer it with or without error checking?
Error checking is best.
I'd agree.
I think the scripts in velvet are fast but will fail horribly with bad input... note there is a simple Biopython script to do this included with velvet already (simple version with no error checking, I have written a more robust version too - it looks like I haven't sent it to Daniel to include in velvet though).
I rolled my own FASTQ paired read interlacer and deinterlacer today, using the Galaxy Python modules in lib/galaxy_utils/. I must say these modules made it quite convenient and efficient to implement error-checking in the (de)interlacing. You can find the scripts here if you're interested: http://bitbucket.org/fangly/galaxy-central I'll make the XML wrappers tomorrow and test them. Hopefully after this is done, my changes can be pulled into the official Galaxy repository.
For the deinterlacer, I previously offered to write something like that for Galaxy and was told to submit it to the Tool Shed initially (although it may be merged into the official repository at some point). See "Divide FASTQ file into paired and unpaired reads" on http://community.g2.bx.psu.edu/ for my tool. I also note you've changed the return behaviour of the Galaxy FASTQ library method get_paired_identifier - that API change could break other parts of Galaxy or 3rd party tools. Looking at that Galaxy lib, perhaps I can offer some of my code for identifying Sanger read pairs and the .f .r suffices to enhance the class fastqJoiner (look like it only does Illumina /1 and /2 right now which I think is too narrow). Peter