A new study warns that contamination from laboratory reagents could be misleading scientists worldwide in their hunt for emerging infectious diseases. Researchers found that silica membranes—commonly used in nucleic acid extraction kits—can harbor parvoviruses and other viral contaminants, creating false virus–host associations in metagenomic sequencing (mNGS) data. These misleading links can affect clinical diagnostics, zoonotic surveillance, and public health responses.
In mNGS analyses of patient samples from multiple regions in China, the team detected dozens of parvovirus sequences that closely matched viruses previously attributed to bats, birds, pangolins, and other animals. Follow-up investigations revealed that these sequences originated not from patients but from the silica columns and certain sampling tubes used in laboratory workflows. Testing across 28 commercial kits showed contamination from 13 different viral families, with Parvoviridae being the most frequent.
To address this widespread problem, the researchers developed the Panoramic Virus Discovery Data Chain (PVDDC)—a high-quality, traceable dataset integrating viral genomic data, laboratory workflows, reagent records, and host associations. Combining big data curation with large language model ChatGPT‐4o, PVDDC standardizes virus–host relationships and enables contamination tracing. They also launched the Parvovirus Database (ParvoDB，http://web3.mgc.ac.cn:8080/parvodb/), a public platform with tools for strain search, human parvovirus records, contamination monitoring, and host–virus network.
By applying this framework, the team found that many parvoviruses linked to humans in public databases lack robust experimental evidence and may instead be artifacts of reagent contamination. Only a small subset—such as human bocavirus and parvovirus B19—has strong, well-supported human associations.
Importantly, the PVDDC framework is not limited to parvoviruses—it can be adapted to investigate and mitigate contamination for a wide range of viral taxa, making it a versatile tool for improving the reliability of mNGS-based pathogen surveillance. The authors encourage researchers worldwide to contribute contamination evidence to ParvoDB, helping to expand its scope and strengthen global capacity to detect, trace, and correctly attribute viruses in clinical and environmental studies.
Experimental study
People
Unveiling Origin Pitfalls in Metagenomic Surveillance for Emerging Infectious Diseases: Parvoviruses as a Model