I have question on standard pipeline arguments:

--input

is mandatory and expects a samplesheet, but when the pipeline is used for just a single sample, it is inconvenient to have to write a samplesheet. I can think of two workarounds: • make

--input

optional and create an alternative arguments that takes a data file directly. Drawback: users can now crash the pipeline by forgetting to provide either

--input

or its alternative. • let

--input

accept other file formats than

csv

. Drawback: deviates from the nf-core template and makes the pipeline harder to maintain. Any thoughts?

Just pass the file directly. Having a CSV is not mandatory. In fact having a samplesheet is a relatively new-ish thing, in the early days all pipelines used globs for that parameter.

Thanks! Accepting CSV format in

--input

is the most straightforward for the user. This is for the #C06TRU74N4D pipeline when the users have just one pair to analyse; in that case the metadata can be hardcoded to target and query (for the input) and I can have optional parameters like

--targetName

and

--queryName

to pass the metadata IDs in case the user wants (as they are used to construct output file names).

Hmm, then probably best to stick with the sample sheet.