Disease gene prediction for molecularly uncharacterized diseases
- Submitting institution
-
Royal Holloway and Bedford New College
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 34135878
- Type
- D - Journal article
- DOI
-
10.1371/journal.pcbi.1007078
- Title of journal
- PLoS Computational Biology
- Article number
- e1007078
- First page
- 1
- Volume
- 15
- Issue
- 7
- ISSN
- 1553-734X
- Open access status
- Compliant
- Month of publication
- July
- Year of publication
- 2019
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
1
- Research group(s)
-
-
- Citation count
- 5
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Published in a top Bioinformatics journals. The first method for the prediction of disease genes for molecularly uncharacterized diseases; hence, particularly important for rare diseases. For molecularly characterized diseases, it outperforms SOTA methods by 14%-65%; and for disease module prediction, it outperforms SOTA methods by 87%-299%. It uses an updatable disease phenotype similarity (developed earlier in my lab) and learns a non-linear transformation to define a prior probability distribution over the genes that mimics the distribution of disease genes in the interactome. Subsequently, a semi-supervised learning method establishes a prioritization ordering for all genes in the interactome.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -