UbiComb: A Hybrid Deep Learning Model for Predicting Plant-Specific Protein Ubiquitylation Sites

Abstract

Protein ubiquitylation is an essential post-translational modification process that performs a critical role in a wide range of biological functions, even a degenerative role in certain diseases, and is consequently used as a promising target for the treatment of various diseases. Owing to the significant role of protein ubiquitylation, these sites can be identified by enzymatic approaches, mass spectrometry analysis, and combinations of multidimensional liquid chromatography and tandem mass spectrometry. However, these large-scale experimental screening techniques are time consuming, expensive, and laborious. To overcome the drawbacks of experimental methods, machine learning and deep learning-based predictors were considered for prediction in a timely and cost-effective manner. In the literature, several computational predictors have been published across species; however, predictors are species-specific because of the unclear patterns in different species. In this study, we proposed a novel approach for predicting plant ubiquitylation sites using a hybrid deep learning model by utilizing convolutional neural network and long short-term memory. The proposed method uses the actual protein sequence and physicochemical properties as inputs to the model and provides more robust predictions. The proposed predictor achieved the best result with accuracy values of 80% and 81% and F-scores of 79% and 82% on the 10-fold cross-validation and an independent dataset, respectively. Moreover, we also compared the testing of the independent dataset with popular ubiquitylation predictors; the results demonstrate that our model significantly outperforms the other methods in prediction classification results.

Genome-wide identification and expression analysis reveals spinach brassinosteroid-signaling kinase (BSK) gene family functions in temperature stress response

“…We performed the prediction of some conserved amino acid sites for protein PTM. The sequence alignment with these reported functional sites in Arabidopsis [ 11 , 18 , 21 , 26 , 27 , 29 , 41 ] and rice [ 16 ], as well as the online tools GPS-SNO 1.0 [ 43 ] and UbiComb [ 44 ], were used for the predictions of phosphorylation, myristoylation, S-nitrosylation, and ubiquitination, respectively (Fig. 3 G).…”

“…The conserved phosphorylation, myristoylation, and reported conserved functional sites in SoBSKs were analyzed by sequence alignment with previous researches in Arabidopsis [ 11 , 18 , 21 , 26 , 27 , 29 , 41 ] and rice [ 16 ]. The S-nitrosylation and ubiquitination sites were predicted by GPS-SNO 1.0 [ 43 ] with default parameters and UbiComb ( http://nsclbio.jbnu.ac.kr/tools/UbiComb/ ) with a threshold value of 0.5 [ 44 ], respectively.…”

Mini-review: Recent advances in post-translational modification site prediction based on deep learning

“…In order to identify the ubiquitination sites in proteins, a myriad of experimental [96] , [97] , [98] and computational methods [99] , [100] , [101] have been developed. In recent years, with the continuous growth in high-throughput experimental data [102] , [103] , [104] , deep learning [105] , [106] , [107] has been increasingly applied to the prediction of ubiquitination. Fu et al proposed a deep learning predictor, DeepUbi [37] , based on CNN.…”

Thirty years of molecular dynamics simulations on posttranslational modifications of proteins

“…Databases and informatic methods cataloguing general PTM information, 54–66 as well as resources depicting experimentally-validated or computationally-predicted sites specific for phosphorylation, 67–88 glycosylation, 80,89–98 and acylation, 80,99–119 among other modification forms, 80,120–132 exist. These resources can be used to identify PTM sites on target proteins for improved molecular modeling.…”