nhKcr: a new bioinformatics tool for predicting crotonylation sites on human nonhistone proteins based on deep learning

Abstract

Lysine crotonylation (Kcr) is a newly discovered type of protein post-translational modification and has been reported to be involved in various pathophysiological processes. High-resolution mass spectrometry is the primary approach for identification of Kcr sites. However, experimental approaches for identifying Kcr sites are often time-consuming and expensive when compared with computational approaches. To date, several predictors for Kcr site prediction have been developed, most of which are capable of predicting crotonylation sites on either histones alone or mixed histone and nonhistone proteins together. These methods exhibit high diversity in their algorithms, encoding schemes, feature selection techniques and performance assessment strategies. However, none of them were designed for predicting Kcr sites on nonhistone proteins. Therefore, it is desirable to develop an effective predictor for identifying Kcr sites from the large amount of nonhistone sequence data. For this purpose, we first provide a comprehensive review on six methods for predicting crotonylation sites. Second, we develop a novel deep learning-based computational framework termed as CNNrgb for Kcr site prediction on nonhistone proteins by integrating different types of features. We benchmark its performance against multiple commonly used machine learning classifiers (including random forest, logitboost, naïve Bayes and logistic regression) by performing both 10-fold cross-validation and independent test. The results show that the proposed CNNrgb framework achieves the best performance with high computational efficiency on large datasets. Moreover, to facilitate users’ efforts to investigate Kcr sites on human nonhistone proteins, we implement an online server called nhKcr and compare it with other existing tools to illustrate the utility and robustness of our method. The nhKcr web server and all the datasets utilized in this study are freely accessible at http://nhKcr.erc.monash.edu/.

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

“…In addition to those discussed, deep learning can also be applied for other PTMs’ predictions, including methylation [110] , S-nitrosylation [111] , succinylation [112] , [113] , malonylation [114] , [115] , S-sulphenylation [116] , [117] , crotonylation [118] , [119] , [120] , [121] , 2- hydroxyisobutyrylation [122] , glutarylation [123] , N-palmitoylation [124] carbonylation [125] , and SUMOylation [126] . In particular, crotonylation prediction has demonstrated highly accurate results based on deep-learning methods.…”

“… 2020 [119] Deep-Kcr Crotonylation Human CNN 10-fold CV 19,928 https://lin-group.cn/server/Deep-Kcr 2020 [120] DeepKcrot Crotonylation Multiple CNN WE 10-fold CV and independent test 10,702/1,265/2,044/5,995 * https://www.bioinfogo.org/deepkcrot . 2021 [121] nhKcr Crotonylation Human CNNrgb 10-fold CV and independent test 180,312 https://nhKcr.erc.monash.edu/ 2021 [118] DeepKhib 2-Hydroxyisobutyrylation Multiple CNN OH 10-fold CV and independent test 18,946/15,444/12,756/19,330/2,098 * https://www.bioinfogo.org/DeepKhib . 2020 [122] DeepGlut Glutarylation Prokaryotes and Eukaryote CNN 10-fold CV 4,572 * https://github.com/urmisen/DeepGlut .…”

Residue–Residue Contact Can Be a Potential Feature for the Prediction of Lysine Crotonylation Sites

“…In 2018, 5,995 sites on 2,120 proteins had first been extracted and released by Liu, K. et al Liu et al (2018) and provided more experimental-verified crotonylated samples in plant Carica papaya L, which filled in the gaps of lacking samples in computational analysis of crotonylation. Based on these Carica papaya L. data, Zhao, Y et al have carried out a prediction on the large dataset, in which the deep learning method has been involved Zhao et al (2020) and more deep learning-related methods were released, such as “DeepKcr” Lv et al (2021) and “nhKcr” Chen Y. Z. et al (2021) .…”

iKcr_CNN: A novel computational tool for imbalance classification of human nonhistone crotonylation sites based on convolutional neural networks with focal loss

“…Later, five traditional ML-based models were developed, including support vector machine (SVM) model by Qiu et al [10] , CKSAAP_CrotSite (SVM) [11] , iKcr-PseEns (ensemble random forest, RF) [12] , LightGBM-CroSite (LightGBM) [13] , and random forest (RF)/SVM classifiers by Wang et al [14] . The remaining five were all built on the frame of deep learning (DL), including iCrotoK-PseAAC [15] , pKcr [16] , Deep-Kcr [17] , DeepKcrot [18] and nhKcr [19] . Throughout these predictors, we can observed that: in terms of datasets, the early tools mainly concentrated on histone or mixed data, whereas the recent works began to study nonhistone proteins due to the enrichment of high-throughout nonhistone data.…”

“…Throughout these predictors, we can observed that: in terms of datasets, the early tools mainly concentrated on histone or mixed data, whereas the recent works began to study nonhistone proteins due to the enrichment of high-throughout nonhistone data. Accordingly, the number of samples sharply increased from 34/90 (the number of positive samples over negative samples) in the first model CrotPred [9] to 15,605/75,111 in the latest model nhKcr [19] , which can effectively guarantee the statistical significance of the constructed models; in terms of protein features, it roughly covered several classical easy-interpreted descriptors (i.e., composition of k-spaced amino acid pairs (CKSAAP), one-hot, enhanced amino acid composition (EAAC), pseudo-amino acid composition (PseAAC), pseudo-position specific scoring matrix (PsePSSM), etc) and deep learning representation embedding methods (i.e., wording embedding (WE)); in terms of algorithms, researchers were more inclined to choose deep learning techniques rather than conventional classifiers; as for model evaluation, recent works strictly completed cross validation and independent tests to get objective and reliable results.…”