Hi all, thanks for the insightful comments this morning. Upon reflection, I think a major limitation in the vocabulary is the overlapping meanings of the term "state" and "model", which is the real origin of the confusion about what an "ensemble" is. I would
like to propose the following resolution:
- Don't use "model" to refer to a particular choice of the model variables/parameters. The model is like a class definition, which can have many instantiations, which would be called "objects" in a programming analogy. You would
never confuse these two in a programming interview.
- Don't use the term "state". Instead, refer to either a "conformational state" or a "model state", defined below
- Conformational State:
A set of spatial coordinates for the modeled system that represent a biological entity. A model representation may include multiple conformational states simultaneously which together make up a
model
state.
- Model State:
A particular instantiation of all the spatial/temporal coordinates defined in the model representation. A model state may encompass multiple
conformational
states, and if so we refer to it as a multi-state model.
By these definitions, we should then
refer to either "conformational ensembles" to mean a distribution of plausible conformational states we might observe our biological object of study in, or "model ensembles" to mean an ensemble over model states, for example an ensemble over pathways (e.g.
a sequence of conformational states) a protein might take from an unfolded to a folded conformational state.
Further clarity could be achieved by replacing "multi-state model" by "multi-conformation model", since "state" is ambiguous.
If it's still confusing, I'll come up with some visual aids for the next meeting. Let me know if you have any comments/thoughts!