iPseU-CNN: Identifying RNA Pseudouridine Sites Using Convolutional Neural Networks

Abstract

Pseudouridine is the most prevalent RNA modification and has been found in both eukaryotes and prokaryotes. Currently, pseudouridine has been demonstrated in several kinds of RNAs, such as small nuclear RNA, rRNA, tRNA, mRNA, and small nucleolar RNA. Therefore, its significance to academic research and drug development is understandable. Through biochemical experiments, the pseudouridine site identification has produced good outcomes, but these lab exploratory methods and biochemical processes are expensive and time consuming. Therefore, it is important to introduce efficient methods for identification of pseudouridine sites. In this study, an intelligent method for pseudouridine sites using the deep-learning approach was developed. The proposed prediction model is called iPseU-CNN (identifying pseudouridine by convolutional neural networks). The existing methods used handcrafted features and machine-learning approaches to identify pseudouridine sites. However, the proposed predictor extracts the features of the pseudouridine sites automatically using a convolution neural network model. The iPseU-CNN model yields better outcomes than the current state-of-the-art models in all evaluation parameters. It is thus highly projected that the iPseU-CNN predictor will become a helpful tool for academic research on pseudouridine site prediction of RNA, as well as in drug discovery.

RF-PseU: A Random Forest Predictor for RNA Pseudouridine Sites

“…The ROC curve and auROC value shown in Figure 2B also demonstrate that the optimized RF models performed better than the optimized SVM models for the same feature spaces. Thus, non-SVM models such as XG-PseU (Liu et al, 2019b), iPseU-CNN (Tahir et al, 2019), and our RF-PseU model might be more suitable for distinguishing pseudouridine sites from nonpseudouridine sites.…”

“…The performance of RF-PseU was also compared with that of state-of-the-art predictors including iRNA-PseU (Chen et al, 2016), PseUI (He et al, 2018), iPseU-CNN (Tahir et al, 2019), and XG-PseU (Liu et al, 2019b). First, we compared the evaluation scores for the three species.…”

“…He et al (2018) presented another model, PseUI, to identify pseudouridine sites in RNA sequences from three species (H. sapiens, S. cerevisiae, and M. musculus); this was an SVM-based model incorporating multiple featureextraction technologies. Tahir et al (2019) used convolutional neural networks to design a new predictor, iPseU-CNN; and Liu et al (2019b) developed the eXtreme gradient boosting (XGboost) method for RNA pseudouridine site prediction (XG-PseU). Cross-validation scores for RNA pseudouridine site identification in the abovementioned three species showed the best accuracy for iPseU-CNN (66.9%) in H. sapiens, whereas XG-PseU and iPseU-CNN had the best accuracy (68.2%) in S. cerevisiae, and XG-PseU was the most accurate (72.0%) in M. musculus.…”

DeepKhib: A Deep-Learning Framework for Lysine 2-Hydroxyisobutyrylation Sites Prediction

“…The number of the negative sites (103,987) is 11 times larger than that of the positive sites. To avoid the potential impact of biased data on model construction, we referred to previous studies and balanced positives and negatives by randomly selecting the same number of negative sites (Huang et al, 2018c;Tahir et al, 2019). These positives and negatives composed the whole human dataset.…”

Improving the Quantification of DNA Sequences Using Evolutionary Information Based on Deep Learning

“…In this regard, deep learning algorithms have produced remarkable results as they are able to learn automatically complex patterns from large datasets. Deep learning has been applied successfully to a wide range of problems such as image and sound processing [3][4][5][6], natural language processing [7], machine translation [8], and various computational biology tasks [9][10][11][12][13][14][15].…”