- Title
- Significance of non-edge priors in gene regulatory network reconstruction
- Creator
- Nair, Ajay; Chetty, Madhu; Wangikar, Pramod
- Date
- 2014
- Type
- Text; Conference paper
- Identifier
- http://researchonline.federation.edu.au/vital/access/HandleResolver/1959.17/157259
- Identifier
- vital:11550
- Identifier
-
https://doi.org/10.1007/978-3-319-12637-1_56
- Identifier
- ISBN:03029743
- Abstract
- It is well known that incorporating prior knowledge improves gene regulatory network reconstruction from data. Two types of prior knowledge can be given for the gene regulatory network inference - known interactions (edge priors) and known absence of interactions (non-edge priors). However, previous studies have focused mainly on edge priors. This paper shows that the edge priors give only limited improvement. Moreover, non-edge priors are crucial for better overall performance and their effect dominates edge priors at larger data samples. The studies are carried out on two real networks and a computationally tractable synthetic network, using Bayesian network framework. Further, a method to obtain large numbers of non-edge priors for real gene regulatory networks is presented. © Springer International Publishing Switzerland 2014.
- Publisher
- Springer
- Relation
- 21st International Conference, ICONIP 2014 Kuching, Malaysia, November 3–6, 2014; published in Neural Information Processing, (Lecture Notes in Computer Science) Vol. 8834 p446-453
- Rights
- Copyright Springer
- Rights
- This metadata is freely available under a CCO license
- Subject
- Bayesian; Data sample; Gene regulatory networks; Incorporating prior knowledge; Network frameworks; Prior knowledge; Real networks; Synthetic networks
- Reviewed
- Hits: 889
- Visitors: 859
- Downloads: 0
| Thumbnail | File | Description | Size | Format |
|---|