Allow `include_altexps` to take a vector in `merge_sce_list`

[jashapiro]

          For the future, part of me wonders if we should allow `include_altexps` to be a vector of `altExps` to include. This would be a change in `scpcaTools`, and for backward compatibility we should allow a `TRUE/FALSE` to mean keep all or keep none, with a vector meaning (implicitly) keep some. 

Originally posted by @jashapiro in AlexsLemonade/scpca-nf#666 (comment)

The idea here is that we could allow

merge_sce_list(sce_list, include_altexp = c("adt")) 

to keep only the adt altExp, even if there were also a cellhash altExp that we wanted to exclude.

merge_sce_list(sce_list, include_altexp = FALSE) 

would still work as before and not include any altExps, and TRUE would include all altExps.

This does not need to be implemented until we think we will have data that actually does contain more than one altExp.