Cascade neural network for predicting epitope on P24 HIV-1

HIV epidemic and high mortality rate encourage scientists and researchers in the world to find effective prevention and treatment method. P24 protein in HIV is an important part of taking the first step of HIV identification. Discovering epitope region in the P24 protein plays an important role in t...

Full description

Saved in:
Bibliographic Details
Main Authors: Rosyda, Miftahurrahma, Adji, Teguh Bharata, Setiawan, Noor Akhmad
Format: Conference Proceeding
Language:English
Subjects:
Amino acids
Coding
Epidemics
Machine learning
Neural networks
Predictions
Proteins
Sensitivity
Sliding
Training
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:HIV epidemic and high mortality rate encourage scientists and researchers in the world to find effective prevention and treatment method. P24 protein in HIV is an important part of taking the first step of HIV identification. Discovering epitope region in the P24 protein plays an important role in the epitope-based vaccine development, it will save many people from death and opportunistic infection that caused by HIV. One of machine learning technique that can be applied in epitope prediction is a neural network. The epitope of HIV contains amino acids that have to be encoded before training using a neural network. The orthogonal encoding was used to encode amino acid sequence, and sliding window method was applied to build training pattern of the amino acid sequence whose different length and slightly number. Optimal result of prediction obtained by doing twice of neural network training, called cascade neural network. Cascade neural network increases the sensitivity approximately 2.78% in each sliding window. The best sensitivity score of 60,49% was obtained with 15 windows size, while the best specificity accuracy was obtained with 9 window size, these are 96,45% and 92,96%.
ISSN:0094-243X
1551-7616
DOI:10.1063/1.4958500