A machine learning-based treatment prediction model using whole genome variants of hepatitis C virus

• DOI: 10.1371/journal.pone.0242028
• Publication Date: 2020-11-05T18:54:28Z
• Authors: Hiroaki Haga; Hidenori Sato; Ayumi Koseki; Takafumi Saito; Kazuo Okumoto; Kyoko Hoshikawa; Tomohiro Katsumi; Kei Mizuno; Taketo Nishina; Yoshiyuki Ueno
• Keywords: Multilayer perceptron; Perceptron; Data set

In recent years, the development of diagnostics using artificial intelligence (AI) has been remarkable. AI algorithms can go beyond human reasoning and build diagnostic models from a number complex combinations. Using next-generation sequencing technology, we identified hepatitis C virus (HCV) variants resistant to directing-acting antivirals (DAA) by whole genome full-length HCV genomes, applied these various machine-learning evaluate preliminary predictive model. genomic RNA was extracted serum 173 patients (109 with subsequent sustained virological response [SVR] 64 without) before DAA treatment. genomes 109 SVR non-SVR were randomly divided into training data set (57 29 non-SVR) validation-data (52 35 non-SVR). The subject nine selected identify optimized combination functional in relation status following therapy. Subsequently, prediction model tested set. most accurate learning method support vector machine (SVM) algorithm (validation accuracy, 0.95; kappa statistic, 0.90; F-value, 0.94). second-most Multi-layer perceptron. Unfortunately, Decision Tree, Naive Bayes could not be fitted our due low accuracy (< 0.8). Conclusively, an rate 95.4% generalization performance evaluation, SVM as best algorithm. Analytical methods based on analysis construction may applicable selection optimal treatment for other viral infections cancer.