RBPLight: a computational tool for discovery of plant-specific RNA-binding proteins using light gradient boosting machine and ensemble of evolutionary features

Abstract

RNA-binding proteins (RBPs) are essential for post-transcriptional gene regulation in eukaryotes, including splicing control, mRNA transport and decay. Thus, accurate identification of RBPs is important to understand gene expression and regulation of cell state. In order to detect RBPs, a number of computational models have been developed. These methods made use of datasets from several eukaryotic species, specifically from mice and humans. Although some models have been tested on Arabidopsis, these techniques fall short of correctly identifying RBPs for other plant species. Therefore, the development of a powerful computational model for identifying plant-specific RBPs is needed. In this study, we presented a novel computational model for locating RBPs in plants. Five deep learning models and ten shallow learning algorithms were utilized for prediction with 20 sequence-derived and 20 evolutionary feature sets. The highest repeated five-fold cross-validation accuracy, 91.24% AU-ROC and 91.91% AU-PRC, was achieved by light gradient boosting machine. While evaluated using an independent dataset, the developed approach achieved 94.00% AU-ROC and 94.50% AU-PRC. The proposed model achieved significantly higher accuracy for predicting plant-specific RBPs as compared to the currently available state-of-art RBP prediction models. Despite the fact that certain models have already been trained and assessed on the model organism Arabidopsis, this is the first comprehensive computer model for the discovery of plant-specific RBPs. The web server RBPLight was also developed, which is publicly accessible at https://iasri-sg.icar.gov.in/rbplight/, for the convenience of researchers to identify RBPs in plants.

RBProkCNN: Deep learning on appropriate contextual evolutionary information for RNA binding protein discovery in prokaryotes

“…Given the limited availability of structural information for RBPs, the majority of existing prediction models relied on sequence-derived features and employed machine learning algorithms for prediction. Notably, among sequence-based features, those derived from PSSM profiles of protein sequences have demonstrated effectiveness in various studies [26] , [27] , [28] , [29] , [30] , [31] , [32] , [33] , [34] , [35] , [36] , [37] . This improvement may stem from the capability of PSSM-based features to capture context-dependent information, conservation patterns, and evolutionary insights.…”

“…Though we initially considered fifteen distinct PSSM-based evolutionary features, two PSSM-derived feature sets, namely KBGM_PSSM and TRGM_PSSM, were selected to harness the complementary information provided by PSSM and other sequence-derived features. While previous RBP prediction models have also incorporated PSSM-derived features like PSSM-400 [27] , [32] , [58] , [59] , BLOSUM62 [28] , [29] , [33] , [34] , [35] , and PSSM-TPC [31] , [60] , the features KBGM_PSSM and TRGM_PSSM have not been explored for RBP prediction in earlier studies.…”

“…In other words, the existing methods were observed achieving less accuracy for predicting prokaryotic-specific RBPs than that of eukaryotic species like humans and mice. The RBPs are known to be highly tissue- and lineage-specific [31] , [38] , [62] , [63] , whereas the existing models have been developed based on protein sequences of several eukaryotic and few prokaryotic species, making them less accurate for predicting prokaryote-specific RBPs. Thus, RBProkCNN is believed to address this gap by providing a model tailored for predicting prokaryote-specific RBPs, underscoring the importance of accurate predictions in unravelling gene expression patterns and enhancing our understanding of prokaryotic gene regulation.…”

“…Feng et al [30] developed iDRBP-ECHF for identifying DBPs and RBPs based on extensible cubic hybrid framework. A plant-specific RBP prediction model, RBPLight was developed by Pradhan et al [31] using evolutionary features derived from PSSM profiles and light gradient boosting method (LightGBM).…”

“…Most of the above stated learning models have been utilized the RBP data from diverse eukaryotic species and a limited number of prokaryotes, resulting in generic models. However, RBPs exhibit specificity not only to individual species but also to lineage-specific families [6] , [22] , [31] , [38] . Therefore, the current generic models may lack the information required for accurate prediction of prokaryote-specific RBPs.…”