第 1 页,共 21 页

Bayesian Validation Subgroup Meeting

Jared Sagendorf

07/21/2023

第 2 页,共 21 页

IHM-Dictionary

Primary data format for archiving of integrative/hybrid structure models in PDB-Dev

Development of a Prototype System for Archiving Integrative/Hybrid Structure Models of Biological Macromolecules

B. Vallat, B. Webb, J. D. Westbrook, A. Sali and H. M. Berman

Structure 2018 Vol. 26 Issue 6 Pages 894-904.e2

Archiving and disseminating integrative structure models

B. Vallat, B. Webb, J. Westbrook, A. Sali and H. M. Berman

Journal of Biomolecular NMR 2019 Vol. 73 Issue 6-7 Pages 385-398

Dictionary browser: https://mmcif.wwpdb.org/dictionaries/mmcif_ihm.dic/Index/

第 3 页,共 21 页

IHM-Dictionary

Contents Overview

model representation

第 4 页,共 21 页

IHM-Dictionary

Extends model representation with coarse-grained spherical beads, three-dimensional Gaussian objects and other geometric primitives such as planes and tori.

第 5 页,共 21 页

Geometric Object Categories

第 6 页,共 21 页

IHM-Dictionary

Supports compositional and conformational heterogeneity and provides representations for multi-state structural models and models related by time or other order

B. Vallat, B. Webb, J. D. Westbrook, A. Sali, and H. M. Berman, Structure, vol. 26, no. 6, pp. 894-904.e2, 2018

第 7 页,共 21 页

IHM-Dictionary

Provides a general representation of spatial restraints and their uncertainties derived from different kinds of biophysical techniques.

第 8 页,共 21 页

Dihedral Restraint

Spatial Restraint

第 9 页,共 21 页

IHM-Dictionary

Provides a generic representation for cross-referencing related data from external resources via stable identifiers, such as accession codes or persistent digital object identifiers.

第 10 页,共 21 页

Model Representation

source: https://github.com/ihmwg/IHM-dictionary/blob/master/dictionary_documentation/documentation.md

Multi-scale

Multi-state

Ensembles

Ordered

第 11 页,共 21 页

Multi-Scale Model Representations

atomic model components

第 12 页,共 21 页

Multi-Scale Model Representations

multi-residue gaussian components

multi-residue spherical components

第 13 页,共 21 页

Models

models are collections of geometric objects representing various components of the system

第 14 页,共 21 页

第 15 页,共 21 页

Model Groups

Model groups act as arbitrary containers for groups of models. Their meaning is based on association with other data tables.

第 16 页,共 21 页

第 17 页,共 21 页

Multi-State Models

States are arbitrarily defined by the modeler and assigned to a model group. Act as meta-data on top of a model group

第 18 页,共 21 页

Ensembles

Ensembles are typically collections of structurally similar models, usually produced by clustering+filtering during the modeling process, but clustering information is not required

Data items in the IHM_ENSEMBLE_INFO category records the details of the model clusters or ensembles obtained after sampling.

第 19 页,共 21 页

Ordered

An ordered “ensemble” is just a directed graph of model groups

Data items in the IHM_ORDERED_ENSEMBLE category records the details of the ensembles ordered by time or other order. Ordered ensembles are described as directed graphs with edges between nodes representing models or model groups.

1

2

5

6

3

4

第 20 页,共 21 页

B. Vallat, B. Webb, J. D. Westbrook, A. Sali, and H. M. Berman, Structure, vol. 26, no. 6, pp. 894-904.e2, 2018

第 21 页,共 21 页

Current Limitations

  • Does not support parametric models
  • Limited ability for recursion/hierarchical models
  • Fixed set of representation types