Back

The Unified Human Virome Database: A toolkit for expanded human virome analysis

Miller, C. J.; Pope, C. E.; Lavitt, M. L.; Caverly, L. J.; LiPuma, J. J.; Penewit, K.; Lewis, J. D.; Salipante, S. J.; Hoffman, L. R.

2026-05-02 microbiology
10.64898/2026.05.01.722327 bioRxiv
Show abstract

Current approaches for computationally analyzing viruses within human microbiomes often rely on databases largely composed of fragmented viral genomes from gastrointestinal samples, limiting identification of viruses exclusively found outside the gastrointestinal tract and analyses requiring high-quality genomes. To address these issues, we created the Unified Human Virome Database (UHVDB), comprising 575,497 high-quality, annotated viral genomes from human gastrointestinal, airway, skin, and urogenital sample metagenomes. We developed an associated toolkit that uses UHVDB to characterize viruses and their potential activity from metagenomes, then applied this toolkit to 1,983 airway sample metagenomes from people with cystic fibrosis. Over half of detected viruses lacked evidence of potential activity and were detected transiently. UHVDB is nearly three times larger than prior viral databases and its ability to identify likely active viruses enables rigorous analysis of viruses from diverse human sample types, expanding the capacity to define virus contributions to health and disease.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.