KELMPSP:基于核极限学习机的假尿苷修饰位点识别

KELMPSP: Pseudouridine Sites Identification Based on Kernel Extreme Learning Machine

假尿苷(ψ)是RNA序列中的一种化学修饰,其在基因转录过程中,由酶的催化作用而形成。它是目前所发现为数最多的一种RNA修饰,并且在正常行使生物学功能方面扮演着重要角色。因此,假尿苷修饰位点的识别是一个非常重要的研究领域。随着RNA序列数据的急速增长,基于机器学习识别假尿苷位点的方法相继提出,但其识别精度有待提高。因此,本文提出了一个新的融合核苷酸化学性质、核苷酸浓度和位置特异性的单核苷酸、双核苷酸、三核苷酸偏好特征的序列编码方式,并基于此编码方式和核极限学习机(Kernel Extreme Learning Machine,KELM)算法,构建了一个新的假尿苷位点预测器,该预测器被称为"KELMPSP"。通过Jackknife测试和独立数据集测试表明,KELMPSP明显优于现有的假尿苷位点预测器。KELMPSP可以通过网站:http://39.105.77.161:8890/KELMPSP进行使用。

Pseudouridine（ ψ） is a chemical modification of the RNA sequence,which is formed by enzymatic catalysis during gene transcription. It is one of the most commonly found RNA modifications and plays an important role in various biological functions. Therefore,the identification of pseudouridine sites is a very important research field. With the rapid growth of RNA sequencing data,machine learning-based methods for identifying pseudouridine sites has been put forward,but their recognition accuracies need to be improved. This paper proposes a new sequence encoding method that combines the nucleotide chemical properties, nucleotide concentration and position-specific mononucleotide,dinucleotide and trinucleotide propensity characteristics. In addition,a new predictor for identifyingpseudouridine sites based on this encoding method and the Kernel Extreme Learning Machine（ KELM）algorithm is built,which is named ＂KELMPSP＂. The experiment performances of Jackknife tests and independent dataset tests show that KELMPSP remarkably outperforms the existing predictors. KELMPSP is available at： http：//39. 105. 77. 161∶ 8890/KELMPSP.