Answers are grounded in this paper and cite their sources.
Introduction Open genomic repositories are critical infrastructure for pathogen surveillance, variant tracking, outbreak reconstruction, and public health decision making, yet many were designed primarily for rapid scientific data sharing rather than adversarial verification. In cybersecurity, Advanced Persistent Threats (APTs) describe long duration, strategically coordinated campaigns against digital systems. This paper extends that logic to genomic surveillance by introducing Advanced Persistent Biological Threats (APBTs): persistent cyber biological campaigns in which malicious actors target sequence data, metadata, reference datasets, and analytical models to degrade surveillance reliability over time. We present, to our knowledge, the first integrated multilayer taxonomy of APBTs, structured around a four phase lifecycle, and examine whether metadata conditions relevant to this threat model are present in an operational genomic repository ecosystem. Methods We audited 100,000 SARS-CoV-2 NCBI BioSample records using 23 checks across five metadata layers: temporal, geographic, host and specimen, provenance and administrative, and technical and analytical. Where selected provenance or technical information was unavailable in BioSample, linked BioProject, SRA, and Nucleotide/GenBank resources were examined before an exposure indicator was assigned. Results Every record triggered at least one APBT relevant metadata exposure indicator. The mean composite exposure score was 5.22 (SD = 2.37), and 51.5% of records were classified as High or Critical. Exposure affected 3.8% of records in the temporal layer, 0.5% in the geographic layer, 52.0% in host and specimen metadata, 74.2% in provenance and administrative metadata, and 100.0% in technical and analytical metadata. Cross repository retrieval recovered substantial information not visible in BioSample alone, particularly for BioProject association, sequence linkage, sequencing platform, and assembly method. Information on who collected the sample was absent from 51.3% of records, while BioProject association remained unavailable for 28.8% after linked NCBI resources were examined. Discussion The findings do not demonstrate deliberate metadata poisoning. However, the structural incompleteness of BioSample metadata creates vulnerabilities that an adversary can exploit to introduce poisoned, misleading, or weakly traceable metadata into genomic repositories. As a result, maliciously altered metadata can appear plausible, remain undetected, and influence downstream genomic analysis and surveillance.
No takes yet. Share an insight, caveat, or question.
Anjum et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: