Hi.
I have used FASTP in Galaxy many times and before I could select single, paired, or paired-end collection, so why can I now only use single or paired-collection? Why can’t both files (forward and reverse) be used independently?
With the paired option I could select forward file and then reverse file without making a paired-collection.
I think this also happens in other programs.
Is it a problem with the tool itself or with Galaxy?
Thanks
Greetings
Hi @H1889
The change to using a collection for paired inputs was intentional. → Missing "paired" input for fastp tool - #5 by bernt-matthias
If you are not sure how to organize your data into a paired-end collection, I posted a little movie last week to demonstrate how to do this (works for any tool that accepts this data shape).
The “why” has to do with how sensitive this tool’s parameters/actions are dependent on paired end read data being input per sample and processed together. A collection folder achieves that and make it much easier to include in a workflow.
I didn’t hide the original datasets but you can to simplify the history. Or, not! This is your choice and you can always unhide collection dataset later on if you want these as individual datasets again. Please see the full post for more details + tutorials and let us know if you have any questions! 
For the older versions that supported individual datasets for paired end data, these are still hosted! Go back about 8 versions to 0.24.0+galaxy3 → FAQ: Changing the tool version. Just keep in mind that tool releases are static for reproducibility reasons. So, if any new functionality has since been added or bug fixes applied, those won’t be back ported to the older releases.
The change log can be reviewed at: Tool form → Options → See in Toolshed → Development repository link (usually a link out to Github) → tool.xml file → then History. For fastp, this would be here: History for tools/fastp/fastp.xml - galaxyproject/tools-iuc · GitHub
Thanks for the explanation, but I do not understand why eliminate that option, is it necessary? Tipically after a cleaning of the reads with FASTP, one run , for example, UNICYCLER that needs the R1 and R2 files not combined into one dataset.
is it possible to separate the files after FASTP?
sorry for the inconvenince.
For the question, you are not alone! Anyone is welcome to comment on the Github ticket in this topic or to start a new one.
But the bigger idea is to use collections, and yes, you can modify collection shapes any way that you want to. If you currently have paired end, then Unzip collection or Flatten collection are common transformations. Both produce a list but are slightly different. Each have a graphic in the tool form Help explaining what they do so you can see the difference.
I understand your point of vew, however, there is no harm in leaving the option of having the two files separately. As I mentioned before, there are many tools that need the two files separately, so allowing the two files separately does not harm the system or users.
I suggest leaving the three options: single, two files, and paired collection.
Best regards
