Welcome @maria.vittoria
Hopefully we can help!
The “mapping over” refers to how collection elements are processed one by one. A longer description is in a hover-over message in the warning. I’ve copied it below. It is pretty technical but the general interpretation is that some tools can consume a List of Pairs collection shape (paired-end) and others expect a simple Flat List shape (single end or a split paired-end).
When a tool consumes a dataset but is run with a collection, the collection maps over the collection. This means instead of just running the tool once - the tool will be run once for each element of the provided collection. Additionally, the outputs of the tool will be collected into a collection that matches the structure of the provided collection. This matching structure means the output collections will have the same element identifiers as the provided collection and they will appear in the same order.
It is easiest to visualize “mapping over” a collection in the context of a tool that consumes a dataset and produces a dataset, but the semantics apply rather naturally to tools that consume collections or produce collections as well.
For instance, consider a tool that consumes a paired collection and produces an output dataset. If a list of paired collections (collection type list:paired) is passed to the tool - it will produce a flat list (collection type list) of output datasets with the same number of elements in the same order as the provided list of paired collections.
In the case of outputs, consider a tool that takes in a dataset and produces a flat list. If this tool is run over a flat list of datasets - that list will be “mapped over” and each element will produce a list. These lists will be gathered together in a nested list structure (collection type list:list) where the outer element count and structure matches that of the input and the inner list for each of those is just the outputs of the tool for the corresponding element of the input.
How this works
Falco, the same as FastQC, is a per-sequence operation, and will expect a Flat List collection type. The input collection will have just List assigned.
After running Flatten Collection on a List of Pairs shape (the collection itself will be labeled with just Pairs), the reads inside of it will have “forward” and a “reverse” added to the element identifiers, and be transformed into a Flat List shape and have just List assigned to the collection.
What to do
Double check that the collection being input has the List annotation, or expand the collection to check. Next, make sure you are selecting the correct collection when running the tool. The message indicates that there is a problem with this? Are you not able to start running the tool?
Later on, you can put all of these steps into a little workflow for reuse!
More help is in this topic, and you could swap in Falco for FastQC in the demonstration workflow. → Quality Control Start Here! multQC issue and guidance? - #2 by jennaj
Let us know how this goes! And if you would like more help, it is difficult to guess more, so you could generate a share link to your history and post it back for troubleshooting feedback! See → https://training.galaxyproject.org/training-material/faqs/galaxy/histories_sharing.html
I’m also running a quick independent test to see if this is working as expected! More soon. 
