XVir: A Transformer-Based Architecture for Identifying Viral Reads from Cancer Samples

TL;DR

XVir is a transformer-based deep learning pipeline designed to accurately detect viral DNA in human tumor samples, addressing challenges posed by viral diversity and improving upon existing methods in accuracy and efficiency.

Contribution

The paper introduces XVir, a novel transformer-based architecture for viral read identification in cancer samples, with improved accuracy and computational efficiency over prior methods.

Findings

XVir achieves high detection accuracy on semi-experimental data.

XVir outperforms state-of-the-art methods in viral detection.

XVir is more compact and less computationally demanding.

Abstract

It is estimated that approximately 15% of cancers worldwide can be linked to viral infections. The viruses that can cause or increase the risk of cancer include human papillomavirus, hepatitis B and C viruses, Epstein-Barr virus, and human immunodeficiency virus, to name a few. The computational analysis of the massive amounts of tumor DNA data, whose collection is enabled by the recent advancements in sequencing technologies, have allowed studies of the potential association between cancers and viral pathogens. However, the high diversity of oncoviral families makes reliable detection of viral DNA difficult and thus, renders such analysis challenging. In this paper, we introduce XVir, a data pipeline that relies on a transformer-based deep learning architecture to reliably identify viral DNA present in human tumors. In particular, XVir is trained on genomic sequencing reads from viral and human genomes and may be used with tumor sequence information to find evidence of viral DNA in human cancers. Results on semi-experimental data demonstrate that XVir is capable of achieving high detection accuracy, generally outperforming state-of-the-art competing methods while being more compact and less computationally demanding.