Rename ihm_ensemble_info category

[brindakv]

The term ensemble is used differently by different communities.

In IHMCIF, we want to consider these as a "collection of models" that individually satisfy the input data within a threshold rather than as ensembles.

Consider renaming ihm_ensemble_info accordingly and update category and item descriptions to reflect the change.

List of released entries with ihm_ensemble_info populated (June 19, 2024):

PDBDEV_00000001
PDBDEV_00000002
PDBDEV_00000003
PDBDEV_00000010
PDBDEV_00000011
PDBDEV_00000012
PDBDEV_00000013
PDBDEV_00000017
PDBDEV_00000018
PDBDEV_00000019
PDBDEV_00000020
PDBDEV_00000021
PDBDEV_00000022
PDBDEV_00000023
PDBDEV_00000025
PDBDEV_00000026
PDBDEV_00000034
PDBDEV_00000035
PDBDEV_00000037
PDBDEV_00000038
PDBDEV_00000041
PDBDEV_00000044
PDBDEV_00000049
PDBDEV_00000050
PDBDEV_00000065
PDBDEV_00000071
PDBDEV_00000072
PDBDEV_00000073
PDBDEV_00000074
PDBDEV_00000077
PDBDEV_00000078
PDBDEV_00000079
PDBDEV_00000081
PDBDEV_00000082
PDBDEV_00000088
PDBDEV_00000090
PDBDEV_00000091
PDBDEV_00000097
PDBDEV_00000112
PDBDEV_00000152
PDBDEV_00000154
PDBDEV_00000155
PDBDEV_00000164
PDBDEV_00000211
PDBDEV_00000213
PDBDEV_00000222
PDBDEV_00000223
PDBDEV_00000224
PDBDEV_00000225
PDBDEV_00000226
PDBDEV_00000227
PDBDEV_00000228
PDBDEV_00000229

[aozalevsky]

@brindakv

individually satisfy the input data within a threshold rather than as ensembles.

this definition contradicts the current model for crosslinking-MS data. we've discussed and agreed on:

A restraint group is considered satisfied, if the condition was met in at least 
one model of the model group/ensemble. 

if it's simply a group of models, shouldn't they be just left as a model group?

[aozalevsky]

as Brinda pointed out, in the IHMCIF paper we wrote:

Second, a model can be multi-state. A set of multiple states can be used to describe a system that exists in a mixture of multiple structural and/or compositional states that collectively satisfy the input information. For example, a sample of enzyme molecules in solution is structurally heterogeneous when it exists in an equilibrium between open and closed states; it is compositionally heterogeneous when it contains enzyme molecules both with and without a ligand.

However, multi-state treatment is an explicit choice in the modeling process. One can model the system as a single state and get a collection of structural models that collectively describe the system (like an IDR ensemble).

Also, in the IHMCIF paper we don't mention ensembles at all.

@jaredsagendorf do you have a good definition for the ensemble?

[aozalevsky]

Also, I just realized that PDBDEV_00000008 is missing from the list. Shouldn't it be a canonical example for the ensemble, where each model has a relatively poor fit to the data but the collection as a whole fits the data well?