Intrinsically disordered proteins (IDPs) are important for a broad range of biological functions and are in- volved in many diseases. An understanding of in- trinsic disorder is key to develop compounds that target IDPs. Experimental characterization of IDPs is hindered by the very fact that they are highly dy- namic. Computational methods that predict disor- der from the amino acid sequence have been pro- posed. Here, we present ADOPT (Attention DisOr- der PredicTor), a new predictor of protein disorder. ADOPT is composed of a self-supervised encoder and a supervised disorder predictor. The former is based on a deep bidirectional transformer, which extracts dense residue-level representations from Facebook’s Evolutionary Scale Modeling library. The latter uses a database of nuclear magnetic reso- nance chemical shifts, constructed to ensure bal- anced amounts of disordered and ordered residues, as a training and a test dataset for protein disor- der. ADOPT predicts whether a protein or a spe- cific region is disordered with better performance than the best existing predictors and faster than most other proposed methods (a few seconds per sequence). We identify the features that are rele- vant for the prediction performance and show that good performance can already be gained with <100 features. ADOPT is available as a stand-alone pack- age at https://github.com/PeptoneLtd/ADOPT and as a web server at https://adopt.peptone.io/.
Istvan Redl, Carlo Fisicaro, Oliver Dutton, Falk Hoffmann, Louie Henderson, Benjamin M.J. Owens, et al. (2022). ADOPT: intrinsic protein disorder prediction through deep bidirectional transformers. NAR GENOMICS AND BIOINFORMATICS, 1, 1-14 [10.1101/2022.05.25.493416].
ADOPT: intrinsic protein disorder prediction through deep bidirectional transformers
Emanuele Paci;
2022
Abstract
Intrinsically disordered proteins (IDPs) are important for a broad range of biological functions and are in- volved in many diseases. An understanding of in- trinsic disorder is key to develop compounds that target IDPs. Experimental characterization of IDPs is hindered by the very fact that they are highly dy- namic. Computational methods that predict disor- der from the amino acid sequence have been pro- posed. Here, we present ADOPT (Attention DisOr- der PredicTor), a new predictor of protein disorder. ADOPT is composed of a self-supervised encoder and a supervised disorder predictor. The former is based on a deep bidirectional transformer, which extracts dense residue-level representations from Facebook’s Evolutionary Scale Modeling library. The latter uses a database of nuclear magnetic reso- nance chemical shifts, constructed to ensure bal- anced amounts of disordered and ordered residues, as a training and a test dataset for protein disor- der. ADOPT predicts whether a protein or a spe- cific region is disordered with better performance than the best existing predictors and faster than most other proposed methods (a few seconds per sequence). We identify the features that are rele- vant for the prediction performance and show that good performance can already be gained with <100 features. ADOPT is available as a stand-alone pack- age at https://github.com/PeptoneLtd/ADOPT and as a web server at https://adopt.peptone.io/.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.