Bonus: From April, I will start a new lab at Osaka University. Virus discovery and phenotypic prediction based on host gene expression will remain a central topic—stay tuned! (16/n)
sites.google.com/view/virus-i...
Finally, I thank my current supervisor, Kei, for providing an environment that made this work possible. This study was primarily supported by JST PRESTO “Pandemic Social Infrastructure” (15/n).
The large-scale analysis of 220k RNA-seq data sets was led by Mai. We also greatly benefited from in-depth discussions with Eddie, Spyros, and Junna (14/n).
This work was driven by three outstanding young researchers: Luca led ISG Profiler, Hiroaki led ISG-VIP, and Kyoko conducted large-scale virus discovery and characterization together with me. I sincerely thank them for their exceptional efforts (13/n).
In summary, ISG-based virus discovery complements conventional virome approaches and provides a scalable solution for comprehensive virus detection in the rapidly expanding landscape of animal RNA-seq data (12/n).
We also uncovered key insights into viral evolution and infection risk, including rat viruses linked to the origins of highly pathogenic parvoviruses in pigs, dogs, and cats, and rodent colonies harboring PRRSV-related viruses at high prevalence (11/n).
These chaphamaparvoviruses were frequently detected in the livers of chickens and wild birds, and infected samples showed viral hepatitis-like transcriptomic signatures, suggesting a potential role in avian hepatitis (10/n).
Notably, geNomad failed to detect viruses of the recently characterized and highly diverse genus Chaphamaparvovirus, whereas our method identified many, including highly divergent viral species (9/n).
Importantly, our approach detected many viral infections missed by the widely used tool geNomad. These were enriched for highly divergent viruses with low sequence similarity to known viruses, highlighting the strength of our strategy for novel virus discovery (8/n).
Using this pre-screening strategy, we analyzed ~40k animal RNA-seq data sets released in 2024 and identified ~2,441 infections across diverse species, including 385 infections by novel viruses (7/n).
ISG-VIP showed a positive prediction rate of ~8% and a recall of ~0.45. When used as a pre-screening step, it reduces the number of samples subjected to computationally intensive virome analysis to 8%, while still capturing ~45% of viral infections (6/n).
In this study, we developed ISG Profiler for rapid ISG quantification from RNA-seq data and ISG-VIP for infection prediction. The framework applies to diverse avian and mammalian species, including those without reference genomes, and processes each sample in ~4 minutes (5/n).
Upon viral infection, interferon-stimulated genes (ISGs) are induced as part of the innate immune response. This response is broadly conserved across vertebrates and observed for diverse viruses, making ISG expression a robust indicator of infection (4/n).
However, conventional virus discovery approaches based on homology searches are computationally expensive and have limited sensitivity for highly divergent viruses. As RNA-seq data continue to grow exponentially, more efficient and scalable strategies are needed (3/n).
Many human infectious diseases arise through zoonotic transmission of animal viruses. To prepare for future outbreaks, it is essential to comprehensively analyze large-scale RNA-seq data from wildlife and livestock to identify previously unknown pathogenic viruses (2/n).
[Please RT] Our new preprint is out!
By rapidly quantifying interferon-stimulated genes (ISGs) expression across 220k RNA-seq data sets from diverse animals, we comprehensively identified “hidden viral infections” in wildlife and livestock (1/n).
www.biorxiv.org/content/10.6...