dorsal/arxiv
View SchemaInformation and Protein Interfaces
| Authors | William W. Chen, Paul J. Choi, Jason E. Donald, Eugene I. Shakhnovich |
|---|---|
| Categories | |
| ArXiv ID | q-bio/0604022 |
| URL | https://arxiv.org/abs/q-bio/0604022 |
Abstract
To confer high specificity and affinity in binding, contacts at interfaces between two interacting macromolecules are expected to exhibit pair preferences for types of atoms or residues. Here we quantify these preferences by measuring the mutual information of contacts for 895 protein-protein interfaces. The information content is significant and is highest at the atomic resolution. A simple phenomenological theory reveals a connection between information at interfaces and the free energy spectrum of association. The connection is presented in the form of a relation between mutual information and the energy gap of the native bound state to off-target bound states. Measurement of information content in designed lattice interfaces show the predicted scaling behavior to the energy gap. Our theory also suggests that mutual information in contacts emerges by a selection mechanism, and that strong selection, or high conservation, of residues should lead to correspondingly high mutual information. Amino acids which contribute more heavily to information content are then expected to be more conserved. We verify this by showing a statistically significant correlation between the conservation of each of the twenty amino acids and their individual contribution to the information content at protein-protein interfaces
{
"annotation_id": "de8461ec-60e6-4fa0-a840-960d69fd3e0c",
"date_created": "2026-03-02T18:01:35.302000Z",
"date_modified": "2026-03-02T18:01:35.302000Z",
"file_hash": "e51705b03b43a91d9fcfe36f17a3d0dad2985ad138e66d3d5a96c9127db68248",
"private": false,
"record": {
"abstract": "To confer high specificity and affinity in binding, contacts at interfaces\nbetween two interacting macromolecules are expected to exhibit pair preferences\nfor types of atoms or residues. Here we quantify these preferences by measuring\nthe mutual information of contacts for 895 protein-protein interfaces. The\ninformation content is significant and is highest at the atomic resolution. A\nsimple phenomenological theory reveals a connection between information at\ninterfaces and the free energy spectrum of association. The connection is\npresented in the form of a relation between mutual information and the energy\ngap of the native bound state to off-target bound states. Measurement of\ninformation content in designed lattice interfaces show the predicted scaling\nbehavior to the energy gap. Our theory also suggests that mutual information in\ncontacts emerges by a selection mechanism, and that strong selection, or high\nconservation, of residues should lead to correspondingly high mutual\ninformation. Amino acids which contribute more heavily to information content\nare then expected to be more conserved. We verify this by showing a\nstatistically significant correlation between the conservation of each of the\ntwenty amino acids and their individual contribution to the information content\nat protein-protein interfaces",
"arxiv_id": "q-bio/0604022",
"authors": [
"William W. Chen",
"Paul J. Choi",
"Jason E. Donald",
"Eugene I. Shakhnovich"
],
"categories": [
"q-bio.BM"
],
"title": "Information and Protein Interfaces",
"url": "https://arxiv.org/abs/q-bio/0604022"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "d3801763-fea0-4604-b9e7-52a9441d63e6",
"id": "arXiv Dataset IDs",
"type": "Model",
"variant": "snapshot-2026-03-01",
"version": "0.1.0"
},
"user_id": 1000002
}