Prediction of serine phosphorylation sites mapping on Schizosaccharomyces Pombe by fusing three encoding schemes with the random forest classifier

Autor: Samme Amena Tasmia, Md. Kaderi Kibria, Khanis Farhana Tuly, Md. Ariful Islam, Mst Shamima Khatun, Md. Mehedi Hasan, Md. Nurul Haque Mollah
Jazyk: angličtina
Rok vydání: 2022
Předmět:
Zdroj: Scientific Reports, Vol 12, Iss 1, Pp 1-14 (2022)
Druh dokumentu: article
ISSN: 2045-2322
DOI: 10.1038/s41598-022-06529-5
Popis: Abstract Serine phosphorylation is one type of protein post-translational modifications (PTMs), which plays an essential role in various cellular processes and disease pathogenesis. Numerous methods are used for the prediction of phosphorylation sites. However, the traditional wet-lab based experimental approaches are time-consuming, laborious, and expensive. In this work, a computational predictor was proposed to predict serine phosphorylation sites mapping on Schizosaccharomyces pombe (SP) by the fusion of three encoding schemes namely k-spaced amino acid pair composition (CKSAAP), binary and amino acid composition (AAC) with the random forest (RF) classifier. So far, the proposed method is firstly developed to predict serine phosphorylation sites for SP. Both the training and independent test performance scores were used to investigate the success of the proposed RF based fusion prediction model compared to others. We also investigated their performances by 5-fold cross-validation (CV). In all cases, it was observed that the recommended predictor achieves the largest scores of true positive rate (TPR), true negative rate (TNR), accuracy (ACC), Mathew coefficient of correlation (MCC), Area under the ROC curve (AUC) and pAUC (partial AUC) at false positive rate (FPR) = 0.20. Thus, the prediction performance as discussed in this paper indicates that the proposed approach may be a beneficial and motivating computational resource for predicting serine phosphorylation sites in the case of Fungi. The online interface of the software for the proposed prediction model is publicly available at http://mollah-bioinformaticslab-stat.ru.ac.bd/PredSPS/ .
Databáze: Directory of Open Access Journals
Externí odkaz:
Nepřihlášeným uživatelům se plný text nezobrazuje