Hello,
I'm trying to write my own wrappers for Cufflinks,Cuffcompare,Cuffdiff, learning mostly by looking at the existing tools. I have a question regarding the multiple input/output implementation in CuffCompare:
In CuffCompare, the XML can accept multiple input files (with a <repeat>), and the <output> section contains the following statements (labels removed for brevity): ========= <outputs> <data format="txt" name="transcripts_accuracy" label="" from_work_dir="cc_output" /> <data format="tabular" name="input1_tmap" label="" from_work_dir="input1.tmap" /> <data format="tabular" name="input1_refmap" label="" from_work_dir="input1.refmap"/>
<data format="tabular" name="input2_tmap" label="" from_work_dir="input2.tmap"> <filter>len( input_files ) > 0</filter> </data> <data format="tabular" name="input2_refmap" label="" from_work_dir="input2.refmap"> <filter>len( input_files ) > 0</filter> </data> <data format="tabular" name="transcripts_tracking" label=""> <filter>len( input_files ) > 0</filter> </data>
<data format="gtf" name="transcripts_combined" label="" from_work_dir="cc_output.combined.gtf"/> </outputs> ======
Assuming cuffcompare generates a TMAP and REFMAP file for each input file, does this mean the wrapper tool retrieves only the first and second input files' TMAP and REFMAP files ? and others are ignored ? Or is there some galaxy black magic python code that iterates multiple inputs and imports them all ?
Thanks, -gordon
Assaf,
Assuming cuffcompare generates a TMAP and REFMAP file for each input file, does this mean the wrapper tool retrieves only the first and second input files' TMAP and REFMAP files ?
Yes, this is the way it works currently. This is clearly problematic, but it meets most users' needs as it enables them to generate a file of combined transcripts--which is needed for Cuffdiff--from many sets of Cufflinks' transcripts. Currently, the Galaxy framework isn't set up to generate multiple outputs for each input; I'm not sure how difficult it will be make these extensions, and it hasn't been a priority as users haven't voiced a need for these files.
Of course, contributions are always welcome if you put something together.
Best, J.
and others are ignored ? Or is there some galaxy black magic python code that iterates multiple inputs and imports them all ?
Thanks for the quick answer,
Jeremy Goecks wrote, On 04/11/2011 12:50 PM:
Assuming cuffcompare generates a TMAP and REFMAP file for each input file, does this mean the wrapper tool retrieves only the first and second input files' TMAP and REFMAP files ?
Yes, this is the way it works currently. This is clearly problematic, but it meets most users' needs as it enables them to generate a file of combined transcripts--which is needed for Cuffdiff--from many sets of Cufflinks' transcripts.
I'm wondering if users are actually looking into those TMAP & REFMAP files, or can they be simply ignored ? Ignoring them would make the tool much simpler (no symlinking patches, etc. ).
galaxy-dev@lists.galaxyproject.org
-
Assaf Gordon -
Jeremy Goecks