Dataset Information

MusiteDeep: a deep-learning framework for general and kinase-specific phosphorylation site prediction.

ABSTRACT:

Motivation

Computational methods for phosphorylation site prediction play important roles in protein function studies and experimental design. Most existing methods are based on feature extraction, which may result in incomplete or biased features. Deep learning as the cutting-edge machine learning method has the ability to automatically discover complex representations of phosphorylation patterns from the raw sequences, and hence it provides a powerful tool for improvement of phosphorylation site prediction.

Results

We present MusiteDeep, the first deep-learning framework for predicting general and kinase-specific phosphorylation sites. MusiteDeep takes raw sequence data as input and uses convolutional neural networks with a novel two-dimensional attention mechanism. It achieves over a 50% relative improvement in the area under the precision-recall curve in general phosphorylation site prediction and obtains competitive results in kinase-specific prediction compared to other well-known tools on the benchmark data.

Availability and implementation

MusiteDeep is provided as an open-source tool available at https://github.com/duolinwang/MusiteDeep.

Contact

xudong@missouri.edu.

Supplementary information

Supplementary data are available at Bioinformatics online.

SUBMITTER: Wang D

PROVIDER: S-EPMC5860086 | biostudies-literature | 2017 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

MusiteDeep: a deep-learning framework for general and kinase-specific phosphorylation site prediction.

Wang Duolin D Zeng Shuai S Xu Chunhui C Qiu Wangren W Liang Yanchun Y Joshi Trupti T Xu Dong D

Bioinformatics (Oxford, England) 20171201 24

<h4>Motivation</h4>Computational methods for phosphorylation site prediction play important roles in protein function studies and experimental design. Most existing methods are based on feature extraction, which may result in incomplete or biased features. Deep learning as the cutting-edge machine learning method has the ability to automatically discover complex representations of phosphorylation patterns from the raw sequences, and hence it provides a powerful tool for improvement of phosphoryl ...[more]

PMID: 29036382

Similar Datasets

Project description:Reversible protein phosphorylation is one of the most pervasive post-translational modifications, regulating diverse cellular processes in various organisms. High throughput experimental studies using mass spectrometry have identified many phosphorylation sites, primarily from eukaryotes. However, the vast majority of phosphorylation sites remain undiscovered, even in well studied systems. Because mass spectrometry-based experimental approaches for identifying phosphorylation events are costly, time-consuming, and biased toward abundant proteins and proteotypic peptides, in silico prediction of phosphorylation sites is potentially a useful alternative strategy for whole proteome annotation. Because of various limitations, current phosphorylation site prediction tools were not well designed for comprehensive assessment of proteomes. Here, we present a novel software tool, Musite, specifically designed for large scale predictions of both general and kinase-specific phosphorylation sites. We collected phosphoproteomics data in multiple organisms from several reliable sources and used them to train prediction models by a comprehensive machine-learning approach that integrates local sequence similarities to known phosphorylation sites, protein disorder scores, and amino acid frequencies. Application of Musite on several proteomes yielded tens of thousands of phosphorylation site predictions at a high stringency level. Cross-validation tests show that Musite achieves some improvement over existing tools in predicting general phosphorylation sites, and it is at least comparable with those for predicting kinase-specific phosphorylation sites. In Musite V1.0, we have trained general prediction models for six organisms and kinase-specific prediction models for 13 kinases or kinase families. Although the current pretrained models were not correlated with any particular cellular conditions, Musite provides a unique functionality for training customized prediction models (including condition-specific models) from users' own data. In addition, with its easily extensible open source application programming interface, Musite is aimed at being an open platform for community-based development of machine learning-based phosphorylation site prediction applications. Musite is available at http://musite.sourceforge.net/.

Project description:MotivationProteins play pivotal roles in biological systems, and precise prediction of their functions is indispensable for practical applications. Despite the surge in protein sequence data facilitated by high-throughput techniques, unraveling the exact functionalities of proteins still demands considerable time and resources. Currently, numerous methods rely on protein sequences for prediction, while methods targeting protein structures are scarce, often employing convolutional neural networks (CNN) or graph convolutional networks (GCNs) individually.ResultsTo address these challenges, our approach starts from protein structures and proposes a method that combines CNN and GCN into a unified framework called the two-model adaptive weight fusion network (TAWFN) for protein function prediction. First, amino acid contact maps and sequences are extracted from the protein structure. Then, the sequence is used to generate one-hot encoded features and deep semantic features. These features, along with the constructed graph, are fed into the adaptive graph convolutional networks (AGCN) module and the multi-layer convolutional neural network (MCNN) module as needed, resulting in preliminary classification outcomes. Finally, the preliminary classification results are inputted into the adaptive weight computation network, where adaptive weights are calculated to fuse the initial predictions from both networks, yielding the final prediction result. To evaluate the effectiveness of our method, experiments were conducted on the PDBset and AFset datasets. For molecular function, biological process, and cellular component tasks, TAWFN achieved area under the precision-recall curve (AUPR) values of 0.718, 0.385, and 0.488 respectively, with corresponding Fmax scores of 0.762, 0.628, and 0.693, and Smin scores of 0.326, 0.483, and 0.454. The experimental results demonstrate that TAWFN exhibits promising performance, outperforming existing methods.Availability and implementationThe TAWFN source code can be found at: https://github.com/ss0830/TAWFN.

Dataset Information

MusiteDeep: a deep-learning framework for general and kinase-specific phosphorylation site prediction.

Motivation

Results

Availability and implementation

Contact

Supplementary information

Publications

MusiteDeep: a deep-learning framework for general and kinase-specific phosphorylation site prediction.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets