Skip to main content

Table 2 Protein structure representations

From: Prop3D: A flexible, Python-based platform for machine learning with protein structural properties and biophysical data

Dimensionality

Representation

Example

1D

Amino Acid Sequence

...MIANE...

2D

Residue•••Residue Graph, with Vertices (Residues) and Edges (Contacts)

figure a

3D

Protein Structure as a 3D Volume, with Atoms as Coordinates in \(\mathbb {R}^{3}\)

figure b

  1. Some fundamentally different types of protein structure representations (reps) are schematized here, arranged by dimensionality of the rep. One can always traverse from higher- to lower-dimensional reps without requiring information, while the reverse is not true. Note that some types of reps are more amenable to encapsulation in simple data structures, e.g. protein sequences as character strings (built-in types for programming languages), and residue•••residue graphs as adjacency matrices (closely related to contact maps). That 3D structures are generally not as ‘cleanly’ representable (in 3D), via available data structures for use in ML workflows, motivates much of Prop3D’s functionality
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us