SNBRFinder: A Sequence-Based Hybrid Algorithm for Enhanced Prediction of Nucleic Acid-Binding Residues

Abstract

Protein-nucleic acid interactions are central to various fundamental biological processes. Automated methods capable of reliably identifying DNA- and RNA-binding residues in protein sequence are assuming ever-increasing importance. The majority of current algorithms rely on feature-based prediction, but their accuracy remains to be further improved. Here we propose a sequence-based hybrid algorithm SNBRFinder (Sequence-based Nucleic acid-Binding Residue Finder) by merging a feature predictor SNBRFinderF and a template predictor SNBRFinderT. SNBRFinderF was established using the support vector machine whose inputs include sequence profile and other complementary sequence descriptors, while SNBRFinderT was implemented with the sequence alignment algorithm based on profile hidden Markov models to capture the weakly homologous template of query sequence. Experimental results show that SNBRFinderF was clearly superior to the commonly used sequence profile-based predictor and SNBRFinderT can achieve comparable performance to the structure-based template methods. Leveraging the complementary relationship between these two predictors, SNBRFinder reasonably improved the performance of both DNA- and RNA-binding residue predictions. More importantly, the sequence-based hybrid prediction reached competitive performance relative to our previous structure-based counterpart. Our extensive and stringent comparisons show that SNBRFinder has obvious advantages over the existing sequence-based prediction algorithms. The value of our algorithm is highlighted by establishing an easy-to-use web server that is freely accessible at http://ibi.hzau.edu.cn/SNBRFinder.

Fig 1. Flowchart of our SNBRFinder algorithm.

SNBRFinder is a sequence-based hybrid prediction algotirhm comprising a feature-based predictor SNBRFinder^F and a template-based predictor SNBRFinder^T. SNBRFinder^F was built using the support vector machine algorithm whose inputs include comprehensive sequence descriptors and SNBRFinder^T was implemented with the sequence alignment algorithm based on profile hidden Markov models.

Fig 2. HHscore distribution of optimal templates for DB312 and RB264.

The HHscore produced by HHblits ranges from 0 to 100% and is used to measure the similarity between the query sequence and its optimal template.

Fig 3. Comparison of our algorithms and existing approaches on three datasets.

(A) DB33, (B) RB49, (C) RB44. In (A) and (B), Accuracy1 = (TP+TN)/(TP+TN+FP+FN) and Accuracy2 = (Sensitivity+Specificity)/2. In (C), the AUC values of SNBRFinder^T, HomPRIP, and PRBR are not provided, because the outputs of these three predictors are binary values. With the exception of SNBRFinder and RNABindRPlus (including the component predictors), the evaluation measures of the other approaches are derived from the recent review articles.

Fig 4. Distribution of the ratio of positive predictions for non-nucleic acid binding and nucleic acid binding proteins.

(A) NB250 annotated by our predictors trained with DB312, (B) NB250 annotated by our predictors trained with RB264. The solid bars represent the prediction results of non-nucleic acid binding sequences, while the hollow bars represent the prediction results of nucleic acid binding sequences.

Fig 5. Snapshots of SNBRFinder web server.

The submission page allows users to input mutiple protein sequences and specify the binding nucleic acid type. When the submitted job is finished, SNBRFinder will demonstrate the prediction results from three perspectives. The first section provides summary information about the query sequence and its optimal template. The second section is graphical representation of the prediction results. The last section includes details about the prediction results such as the outputs from our three predictors.