Explore the idea of "suggested filtering" colData and rowData fields #134
Description
This is getting filed here because it is of an exploratory nature. Implementation, should things move forward, will likely happen in scpcaTools.
Context/idea
Related to #105
In the filtered RDS we provide to users, we can perform minimal filtering (e.g., remove genes that are not detected) while allowing folks to decide for themselves how aggressive (or not!) they'd like to be in filtering. The way we've decided to facilitate this is by adding relevant metadata (e.g., miQC probability) to the filtered RDS objects. We have discussed adding additional boolean fields that let users know whether we would filter something out or not, without actually removing rows and columns ourselves. (We were tentatively calling this ccdl_suggests but may be able to do better naming.)
Next steps
We need some idea of what our suggestions to be, so I'm going to argue that the first step is to come up with a game plan for how we would gather additional information and support for our suggestions. Example question I have: What variety of samples would we need to process and explore to feel confident about those recommendations?