BrainPepPass: A Framework Based on Supervised Dimensionality Reduction for Predicting Blood-Brain Barrier-Penetrating Peptides

被引：1

作者：

de Oliveira, Ewerton Cristhian Lima ^{[1
,2
]}

Hirmz, Hannah ^{[3
]}

Wynendaele, Evelien ^{[3
]}

Feio, Juliana Auzier Seixas ^{[1
]}

Moreira, Igor Matheus Souza ^{[1
]}

da Costa, Kaue Santana ^{[5
]}

Lima, Anderson H. ^{[4
]}

De Spiegeleer, Bart ^{[3
]}

de Sales Jr, Claudomiro de Souza ^{[1
]}

机构：

[1] Univ Fed Para, Lab Inteligencia Computac & Pesquisa Operac, Inst Tecnol, BR-66075110 Belem, Para, Brazil

[2] Inst Tecnol Vale, BR-66055090 Belem, Para, Brazil

[3] Univ Ghent, Fac Pharmaceut Sci, Drug Qual & Registrat DruQuaR Grp, B-9000 Ghent, Belgium

[4] Univ Fed Para, Lab Planejamento & Desenvolvimento Farmacos, Inst Ciencias Exatas & Nat, BR-66075110 Belem, Para, Brazil

[5] Univ Fed Oeste Para, Lab Simulacao Computac, BR-68040255 Santarem, Para, Brazil

来源：

JOURNAL OF CHEMICAL INFORMATION AND MODELING | 2023年 / 64卷 / 07期

关键词：

CENTRAL-NERVOUS-SYSTEM; IN-SILICO PREDICTION; DRUG DISCOVERY; COMPUTATIONAL APPROACH; INSULIN TRANSPORT; SHUTTLE PEPTIDES; QSAR MODELS; PERMEABILITY; PERMEATION; VITRO;

D O I：

10.1021/acs.jcim.3c00951

中图分类号：

R914 [药物化学];

学科分类号：

100701 ;

摘要：

Peptides that pass through the blood-brain barrier (BBB) not only are implicated in brain-related pathologies but also are promising therapeutic tools for treating brain diseases, e.g., as shuttles carrying active medicines across the BBB. Computational prediction of BBB-penetrating peptides (B3PPs) has emerged as an interesting approach because of its ability to screen large peptide libraries in a cost-effective manner. In this study, we present BrainPepPass, a machine learning (ML) framework that utilizes supervised manifold dimensionality reduction and extreme gradient boosting (XGB) algorithms to predict natural and chemically modified B3PPs. The results indicate that the proposed tool outperforms other classifiers, with average accuracies exceeding 94% and 98% in 10-fold cross-validation and leave-one-out cross-validation (LOOCV), respectively. In addition, accuracy values ranging from 45% to 97.05% were achieved in the independent tests. The BrainPepPass tool is available in a public repository for academic use (https://github.com/ewerton-cristhian/BrainPepPass).

引用

页码：2368 / 2382

页数：15