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