Deep learning extends de novo protein modelling coverage of genomes using iteratively predicted structural constraints
- Submitting institution
-
University College London
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 14705
- Type
- D - Journal article
- DOI
-
10.1038/s41467-019-11994-0
- Title of journal
- Nature Communications
- Article number
- 3977
- First page
- -
- Volume
- 10
- Issue
- 1
- ISSN
- 2041-1723
- Open access status
- Compliant
- Month of publication
- September
- 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
- Yes
- Number of additional authors
-
2
- Research group(s)
-
-
- Citation count
- 19
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper was first to exploit the recent developments in deep learning-based covariation modelling for a practical life sciences application. DMPfold is conceptually similar to DeepMind’s AlphaFold, though developed independently and published first. DMPfold is open source software, and employs different neural network architectures and a novel data augmentation strategy. A key aspect of this paper is that we used DMPfold to build models for the “dark proteome” i.e. proteins for which no structure was known. The results clearly show that DMPfold was representative of the state of the art compared to previous studies even just 2 years old.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -