ViraMiner: Deep learning on raw DNA sequences for identifying viral genomes in human samples.

Ardi Tampuu,Joakim Dillner,Raul Vicente,Zurab Bzhalava,Ulrich Melcher

doi:10.1371/journal.pone.0222271

Abstract

Despite its clinical importance, detection of highly divergent or yet unknown viruses is a major challenge. When human samples are sequenced, conventional alignments classify many assembled contigs as "unknown" since many of the sequences are not similar to known genomes. In this work, we developed ViraMiner, a deep learning-based method to identify viruses in various human biospecimens. ViraMiner contains two branches of Convolutional Neural Networks designed to detect both patterns and pattern-frequencies on raw metagenomics contigs. The training dataset included sequences obtained from 19 metagenomic experiments which were analyzed and labeled by BLAST. The model achieves significantly improved accuracy compared to other machine learning methods for viral genome classification. Using 300 bp contigs ViraMiner achieves 0.923 area under the ROC curve. To our knowledge, this is the first machine learning methodology that can detect the presence of viral sequences among raw metagenomic contigs from diverse human samples. We suggest that the proposed model captures different types of information of genome composition, and can be used as a recommendation system to further investigate sequences labeled as "unknown" by conventional alignment methods. Exploring these highly-divergent viruses, in turn, can enhance our knowledge of infectious causes of diseases.

Highlights

The human virome is the collection of all viruses that reside in and on the human body
In this work we developed ViraMiner, a Convolutional Neural Networks (CNN)-based method for detecting viral contigs in human metagenomic datasets
The model achieves 0.923 test area under the Receiver Operating Characteristic (ROC) curve. This is a significant improvement compared to a previous tool ([28], reaching 0.79 AUROC) that was designed using relative synonymous codon usage (RSCU) and trained on data originating from the same metagenomic experiments

Summary

Introduction

The human virome is the collection of all viruses that reside in and on the human body. Many different viruses are present in human samples and their composition appears to be different in diseased individuals [1, 2]. Its full impact on human health is not fully understood [3, 4] and the detection and classification of human viruses represents a major challenge. Current metagenomic studies detect many novel viruses, which indicates that only a small part of human viruses has been discovered and many others are yet to be reported [5,6,7,8,9,10]. Studies report epidemiological indications that there may exist undiscovered pathogens.

Methods

Results

Conclusion