I have an srf file that I'm trying to unpack into Bustard directory, which I then will run Gerald on. I ran srf2illumina for this purpose ( version 1.11.3 ), and it produced a bunch of intensity/noise/etc files, plus a few files named "8_ILLUMNINA_GA*" files. I then moved / created a dummy directory structure so that GERALD.pl would be happy.
I then ran GERALD.pl ( from GA pipeline version 1.0rc4 ), which produced a makefile with no apparent erros, then ran Make in that directory. This failed mid-way through because it couldn't find the xml params file at the jerboa.pl stage. I then ran jerboa.pl on the 8_ILLUMINA_GA_PARAMS.txt file, which is an xml file, and disovered that the file format is mismatched: for instance, in the 8_ILLUMINA_*.file, the Instrument tag falls under "Run->Run Parameters->" while in jerboa.pl, it is looking for it under "run" ( note, the case differs as well ).
( The directory structure I created looks like this:
Is there a versioning issue at work here? What GERALD should I use to match with the given srf2illumina, or vice versa? Plus, am I using this tool in some fundamentally wrong way?
The srf2illumina tool was initially intended simply as a debugging and/or verification tool to check that the data stored in the SRF files is valid and matching (approximately - we truncate the digit after the decimal point) the input run folder.
It's been fleshed out a bit since then to a more stable tool, but it still shows some inadequacies. It makes no attempt to create files in appropriate directories. You'll probably need to run Goat or similar to construct all the directories and Makefiles and then to move the srf2illumina files into their appropriate places. Some will need renaming too as the original filenames are .params, s_something_matrix.txt, etc instead of the all-caps names given by srf2illumina.
Someone may have already done this I guess, but I'm not aware of it if so.