dorsal/arxiv
View SchemaConserved network motifs allow protein-protein interaction prediction
| Authors | Istvan Albert, Reka Albert |
|---|---|
| Categories | |
| ArXiv ID | q-bio/0406042 |
| URL | https://arxiv.org/abs/q-bio/0406042 |
Abstract
High-throughput protein interaction detection methods are strongly affected by false positive and false negative results. Focused experiments are needed to complement the large-scale methods by validating previously detected interactions but it is often difficult to decide which proteins to probe as interaction partners. Developing reliable computational methods assisting this decision process is a pressing need in bioinformatics. We show that we can use the conserved properties of the protein network to identify and validate interaction candidates. We apply a number of machine learning algorithms to the protein connectivity information and achieve a surprisingly good overall performance in predicting interacting proteins. Using a 'leave-one-out' approach we find average success rates between 20-50% for predicting the correct interaction partner of a protein. We demonstrate that the success of these methods is based on the presence of conserved interaction motifs within the network. A reference implementation and a table with candidate interacting partners for each yeast protein are available at http://www.protsuggest.org
{
"annotation_id": "49494f26-19b8-48b7-92e3-c96d04d6b070",
"date_created": "2026-03-02T18:01:32.232000Z",
"date_modified": "2026-03-02T18:01:32.232000Z",
"file_hash": "7f1a2adef0a37d2aa57e36359c40ea7cac1c6d0e8acbca28b962886e0d8c9735",
"private": false,
"record": {
"abstract": "High-throughput protein interaction detection methods are strongly affected\nby false positive and false negative results. Focused experiments are needed to\ncomplement the large-scale methods by validating previously detected\ninteractions but it is often difficult to decide which proteins to probe as\ninteraction partners. Developing reliable computational methods assisting this\ndecision process is a pressing need in bioinformatics. We show that we can use\nthe conserved properties of the protein network to identify and validate\ninteraction candidates. We apply a number of machine learning algorithms to the\nprotein connectivity information and achieve a surprisingly good overall\nperformance in predicting interacting proteins. Using a \u0027leave-one-out\u0027\napproach we find average success rates between 20-50% for predicting the\ncorrect interaction partner of a protein. We demonstrate that the success of\nthese methods is based on the presence of conserved interaction motifs within\nthe network. A reference implementation and a table with candidate interacting\npartners for each yeast protein are available at http://www.protsuggest.org",
"arxiv_id": "q-bio/0406042",
"authors": [
"Istvan Albert",
"Reka Albert"
],
"categories": [
"q-bio.MN",
"q-bio.GN"
],
"title": "Conserved network motifs allow protein-protein interaction prediction",
"url": "https://arxiv.org/abs/q-bio/0406042"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "0ba42999-6517-45d4-899e-003265659a51",
"id": "arXiv Dataset IDs",
"type": "Model",
"variant": "snapshot-2026-03-01",
"version": "0.1.0"
},
"user_id": 1000002
}