Stephanie Goya

In November 2025, the first global H5N5 infection was reported in a person in Washington State. Previously, these viruses had only been detected in seabirds and shorebirds near the Atlantic Ocean, suggesting potentially distinct ecology from dominant H5N1 strains. Using 354 global H5N5 genomes, we show that H5N5 viruses have been carried by high latitude seabirds and shorebirds across the Atlantic at least 9 times, with introductions more frequent in the fall and spring. We reconstruct a plausible path of cross-continental movement via Arctic or sub-Arctic regions, followed by a year of cryptic circulation prior to human spillover. These data show that H5N5 viruses have undergone far more intercontinental movements than previously recognized, and exhibit host and geographic patterns that are fundamentally distinct from dominant H5N1 lineages. The distinct host ecology, geography, and movement, paired with the capacity to infect a person, suggest that enhanced surveillance may be warranted.

Rhinoviruses and enteroviruses are among the most frequently detected respiratory viruses, yet routine diagnostics do not report species or genotype, and standard VP1 genotyping cannot resolve transmission chains or recombination. Recovering sequence beyond VP1 currently requires probe-based enrichment, deep metagenomic sequencing, or several overlapping amplicons. We developed a single-amplicon PCR and sequencing workflow that generates coding-complete genomes from both enteroviruses and rhinoviruses in a single reaction. We evaluated the protocol on 140 rhinovirus-positive respiratory specimens and a panel of 41 isolates spanning 26 enterovirus A-D types, comparing it with hybridization capture sequencing and relating its performance to viral load measured by RT-qPCR and droplet digital RT-PCR. Single-amplicon sequencing recovered 116 high-quality rhinovirus genomes, across 57 rhinovirus genotypes, compared with 113 from hybridization capture. Single-amplicon sequencing was associated with higher fidelity and more uniform sequencing depth compared to hybridization capture. The same primer configuration also recovered two unexpected coxsackieviruses (A6 and A16) from rhinovirus-positive clinical specimens, and the expected polyprotein-coding sequences from all 40 enterovirus isolates tested. Recovery increased with viral abundance, and a visible amplicon band predicted sequencing success. Single-amplicon sequencing reagents cost approximately three-fold less per sample than hybridization capture ($30 versus $85) and required one-third as much time. These findings support single-amplicon sequencing as a simple, cost-effective approach for genomic surveillance of enteroviruses and rhinoviruses from clinical specimens. Importance: Routine respiratory virus tests can detect rhinoviruses and enteroviruses but seldom identify the circulating viral lineage, while VP1 genotyping of rhinoviruses and enteroviruses cannot adequately resolve transmission or recombination. Genome sequencing can provide this information, but existing approaches may require extensive sequencing, probe-based enrichment, or multiple overlapping PCR products. We describe a single-round, single-amplicon method that recovers the polyprotein-coding region across both rhinoviruses and enteroviruses. The method produced highly uniform coverage and reproducible consensus sequences at approximately one-third the price of hybridization capture, and it includes a simple pre-sequencing quality check. This streamlined workflow can help make enterovirus and rhinovirus genomic surveillance more accessible to clinical and public health laboratories.

This protocol amplifies the coding-complete genome of rhinoviruses and enteroviruses (approximately 7 kb) from extracted RNA in a single RT-PCR, using a 5′ UTR forward primer (f448) and a reverse primer that targets the tag introduced during reverse transcription with T30TAG, and prepares the product for Illumina sequencing. A visible band on an agarose gel before library preparation indicates successful amplification. In 140 rhinovirus-positive respiratory specimens, the protocol yielded genomic data from 84% of specimens, including six co-infections with two rhinovirus genotypes (of the same or different species) and one co-infection of rhinovirus with coxsackievirus. The protocol also recovered high-quality genomes of 26 enterovirus A–D types in a panel of 41 isolates.

Motivation: Unique molecular identifiers (UMIs) are widely used in next-generation sequencing to enable accurate molecular counting and error correction. However, challenges remain in accurately collapsing UMI clusters, especially when read counts are low or sparse read clusters arise from barcode sequencing errors. Results: We present RUMINA, a Rust-based pipeline for UMI-aware deduplication and error correction, optimized for both amplicon and shotgun sequencing. RUMINA supports multiple UMI cluster strategies, alongside majority-rule read selection independent of mapping quality, as well as discrete handling of 1–2 read clusters, paired-end merging, and read-length stratification. Benchmarking using simulated HIV population sequencing data and real-world iCLIP and TCR datasets showed that RUMINA improves ultra-low frequency SNV detection (0.01%–1%), reduces false positives, enhances reproducibility, and processes sequencing data up to 10-fold faster than existing tools. By integrating UMI- and sequence-level correction in a high-performance framework, RUMINA offers a fast, scalable, and robust solution for UMI-enabled sequencing workflows. Availability and implementation: RUMINA is implemented in Rust and distributed as open-source code and precompiled binaries. Source code and installation instructions are available at https://github.com/greninger-lab/rumina. Documentation associated with this manuscript is available at https://github.com/greninger-lab/rumina_paper.

Anelloviruses are highly diverse, ubiquitous single-stranded DNA viruses whose replication, cellular reservoirs, and mechanisms of persistence remain poorly understood. Here, we identify and characterize a Alphatorquevirus homin 29 genome stably integrated into an rDNA locus of human chromosome 21 in the acute myeloid leukemia cell line SKNO-1. Large-scale mining of NCBI Sequence Read Archive datasets revealed unusually high anellovirus k-mer abundance specifically in sequencing runs of the SKNO-1 cell line from multiple laboratories. De novo assembly of RNA-seq data recovered a 3.25 kb viral genome, and long-read PacBio sequencing confirmed its integration within the RNA45SN2 gene on human chromosome 21. Digital droplet PCR (ddPCR) quantified ∼0.5 viral genomes per cell, and RT-ddPCR detected transcripts from ORF1 and viral terminal repeats, consistent with transcription of the integrant. The ORF1 coding sequence encoded for a truncated but structurally conserved capsid protein retaining jelly roll and P-domain features but lacking the C-terminal domain. Analysis of public ChIP-seq data from the SKNO-1 cell line demonstrated broad occupancy of the BRD4 transcriptional coactivator across the viral genome along with focal enrichment of hematopoietic ETS family transcription factors within the integrant’s ∼300 bp untranslated region upstream of viral open reading frames. Together, these findings demonstrate stable integration, chromatin accommodation, and transcriptional maintenance of an Alphatorquevirus in a human leukemia cell line, providing a unique model to study host tolerance and transcriptional regulation of anellovirus DNA and informing our understanding of anellovirus hematopoietic cell tropism. IMPORTANCE: Anelloviruses are a curious group of highly genetically diverse single-stranded DNA viruses that have been found ubiquitously among humans and are hypothesized to be a potential commensal human virus. Their omnipresence has been matched only by the dearth of our understanding of their basic biological processes – how they persist, how they replicate, and how they spread. Here we identify a naturally integrated Alphatorquevirus genome in the widely used AML cell line SKNO-1 and show that it is stably maintained, transcribed, and embedded within a rDNA locus on human chromosome 21. This discovery highlights the ability to anelloviruses to integrate into human DNA, generates hypotheses around transcription factors used in anellovirus gene transcription, and suggests anelloviruses can persist in myeloblasts.

Human parainfluenza virus 2 (HPIV-2) and human parainfluenza virus 4 (HPIV-4) are significant but underappreciated respiratory pathogens, particularly among high-risk populations including children, the elderly, and immunocompromised individuals. In this study, we sequenced 101 HPIV-2 and HPIV-4 genomes from respiratory samples collected in western Washington State and performed comprehensive evolutionary analyses using both new and publicly available sequences. Phylogenetic and phylodynamic analyses revealed that both HPIV-2 and HPIV-4 evolve at significantly faster rates compared to the mumps virus, a reference human orthorubulavirus. Notably, while HPIV-2 demonstrated the highest evolutionary rates in the surface glycoprotein HN, consistent with humoral immune-driven selection, the innate immune antagonist V/P gene evolved fastest in HPIV-4. We identified a hypervariable region within the HPIV-4V/P protein (residues 35 to 75), which structural modeling placed in a loop overlapping a known interferon antagonism domain in other paramyxovirus V proteins, though HPIV-4 is functionally incompetent in this activity. Expanded phylogenetic analysis across the Paramyxoviridae family uncovered a striking evolutionary discordance: while the HN glycoprotein and L polymerase of HPIV-4 and its 2 closest bat-derived viruses clustered within the Orthorubulavirus genus, their nucleoprotein (N), phosphoprotein (P), matrix (M), and fusion (F) proteins formed a distinct lineage outside the Rubulavirinae subfamily. Together, these findings highlight the distinct evolutionary trajectories of HPIV-2 and HPIV-4, raise hypotheses around complex Paramyxoviridae zoonotic events including recombination-like patterns, and demonstrate limitations of current L protein-based taxonomic classification schemes.

Accurate annotation of viral genomes is essential for reliable downstream analysis and public data sharing. While NCBI’s Viral Annotation DefineR (VADR) pipeline provides standardized annotation and quality control, it only supports six viral groups to date. Here, we developed and validated 12 new reference sequence-based VADR models targeting key human respiratory viruses: measles virus, mumps virus, rubella virus, human metapneumovirus, human parainfluenza virus types 1–4, and seasonal coronaviruses (229E, NL63, OC43, HKU1). Model construction was guided by a comprehensive analysis of intra-species genomic and phylogenetic diversity, enabling the development of genotype-specific models associated with reference genomes that defined expected genome structure and annotation. Models were trained on 5,327 publicly available complete viral genomes and tested on 372 viral genomes not yet submitted to GenBank. VADR passed 96.3% of publicly available viral genomes and 98.1% of viral genomes not in the training set, correctly identifying overlapping ORFs, mature peptides, and transcriptional slippage as well as genome misassemblies. VADR detected novel viral biology including the first reported HCoV-OC43 NS2 knockout in a human infection and novel G and SH coding sequence lengths in human metapneumovirus. These VADR models are publicly available and are used by NCBI curators as part of the GenBank submission pipeline, supporting high-quality, scalable viral genome annotation for research and public health.

Human respiratory syncytial virus (HRSV) is a common cause of lower respiratory tract infections globally, and changes in viral epidemiology have been observed in many jurisdictions following the coronavirus disease 2019 (COVID-19) pandemic. Newly licensed vaccines and monoclonal antibodies are anticipated to alleviate the burden on healthcare systems, though such interventions may exert selective pressures on viral evolution. To evaluate the diversity of HRSV in Ireland pre- and post-COVID-19 pandemic, whole-genome sequencing was performed on HRSV-A (n=123) and -B (n=110) samples collected from community and hospitalized cases, during three HRSV seasons between 2021 and 2024. Additionally, G gene sequences, from HRSV-A (n=141) and -B (n=141), collected in the 2015–2019 period were examined. Lineages were assigned by phylogenetic analyses including reference lineages. Phylogenetic trees inferred with the G gene and whole genomes were consistent. Changes in the prevalence of certain lineages post-COVID-19 reflected the impact of non-pharmaceutical interventions (NPIs) introduced to reduce severe acute respiratory syndrome coronavirus 2 transmission, with A.D.1 and A.D.5 the dominant HRSV-A lineages and B.D.E.1 the most prevalent HRSV-B lineage. Similar trends were observed in HRSV lineages circulating across Europe during this time. The emergence of a new lineage was identified as a descendant from A.D.1, with eight distinctive substitutions in proteins G, F and L. Other circulating lineages with aa substitutions were observed in the F glycoprotein, which could impact nirsevimab binding. We provide the first comprehensive analysis of HRSV genomic diversity and evolution in Ireland over the last decade and the impact of the NPIs introduced during the COVID-19 pandemic. This study provides a foundation for future public health surveillance employing pathogen genomics to enable an evidence-based assessment of the impact of pharmaceutical interventions on HRSV evolution and disease severity.

Background: In December 2024, human metapneumovirus (hMPV) gained global attention amid rising cases in Chinese hospitals, prompting a World Health Organization (WHO) statement indicating case numbers remained within expected ranges. To assess whether a variant of public health concern was emerging and to examine global hMPV phylodynamics, we conducted a genomic surveillance study in western Washington State. Study design: We sequenced hMPV genomes from inpatient and outpatient samples collected between 2021–2025 in western Washington State and constructed phylogenomic and phylodynamic trees. Results: We obtained 60 hMPV-A and 39 hMPV-B genomes, including 13 from November 2024–January 2025. Following COVID-19 pandemic disruptions, hMPV seasonality returned to typical patterns after 2023. Genomic analysis showed hMPV-A predominance since 2022–23, with co-circulation of A2b1, A2b2, B1, and B2 sublineages. The A2b2 sublineage was most prevalent and all genomes carried a 111-nt G gene duplication. Phylogenetic evidence suggests the 111-nt variant evolved from a prior 180-nt duplication via a 69-nt deletion, rather than through independent duplication events. Most sublineages showed multiple co-circulating clades, except A2b1. Phylodynamics showed recovery of global diversity after pandemic-related declines and a higher evolutionary rate in hMPV-A compared to hMPV-B. No distinct evolutionary features of heightened concern were detected. Conclusions: Despite recent concerns, our findings indicate that hMPV circulation in the USA follows expected seasonal patterns, with ongoing introductions of diverse viral variants from preexisting sublineages rather than emergence of a novel variant. Continued genomic surveillance is essential, particularly as vaccine development progresses. Highlights: hMPV seasonality in Washington State normalized by 2023 after pandemic-related disruption. Recent hMPV genomes show multiple introductions without evidence of a novel variant of concern. A2b2-111, the dominant sublineage, likely evolved from A2b2-180 variant via 69-nt G gene deletion. hMPV-A evolves faster than hMPV-B; global diversity declined in 2020 and rebounded post-2021.