Nanopore Adaptive Sampling for Targeted Pathogen Sequencing: Principles, Workflows, and Applications in Veterinary Diagnostics

Introduction

Rapid and accurate identification of pathogens in veterinary specimens is a cornerstone of infectious disease control in animal populations. Metagenomic sequencing offers an unbiased approach to pathogen detection but often suffers from low analytical sensitivity when target nucleic acid is scarce relative to host background [1, 2, 3]. Targeted enrichment methods, such as hybrid capture or amplicon-based approaches, increase sensitivity but require prior knowledge of target sequences and incur additional library preparation steps [4, 5]. Nanopore adaptive sampling, a software-driven selection method integrated into the real-time basecalling process, provides an alternative strategy that enriches for user-defined genomic regions without physical capture or amplification [6, 7, 8]. This article reviews the biophysical and computational principles of adaptive sampling, its implementation in veterinary pathogen sequencing, and its current performance characteristics as reported in the peer-reviewed literature. The discussion is confined to animal health contexts, drawing on studies of ticks, carnivores, poultry, and livestock pathogens.

Biophysical and Algorithmic Foundations of Adaptive Sampling

Nanopore sequencing relies on the translocation of single-stranded DNA or RNA molecules through a protein nanopore embedded in an electrically resistant membrane. As the polynucleotide traverses the pore, it modulates the ionic current in a sequence-dependent manner, yielding raw electrical signals that are converted into nucleotide reads through basecalling algorithms [9]. Adaptive sampling exploits the real-time nature of this process by making dynamic enrichment or depletion decisions during sequencing [6, 10].

The workflow proceeds as follows. After a read begins translocation, the sequencing device performs an initial "squiggle" analysis on the first few hundred milliseconds of signal. This partial alignment against a reference index (typically a target genome or a set of target sequences) determines whether the read originates from a region of interest [6, 11]. If the read matches a target, the voltage across the pore is maintained, allowing the full read to be sequenced. If the read matches a non-target sequence (e.g., host DNA), the voltage is reversed, ejecting the molecule from the pore and freeing the channel for a subsequent molecule [7, 8]. This rejection occurs within tens of milliseconds, minimizing wasted sequencing capacity on background DNA.

The core algorithmic components include a fast approximate aligner (often based on minimizers) that operates on raw signal or early basecalled data, and a decision module that controls pore voltage [6, 12]. The reference index can be constructed from complete genome sequences, specific marker genes, or even synthetic probes designed by generative adversarial networks [6]. Consequently, adaptive sampling enables enrichment of target molecules by factors ranging from 5-fold to 50-fold depending on the relative abundance of target and background, the length of the target regions, and the complexity of the sample [13, 10].

Comparison with Other Enrichment Strategies

Several methods exist for enriching target nucleic acid before or during sequencing. Hybrid capture uses biotinylated probes complementary to target sequences, followed by streptavidin bead pull-down [4, 14, 5]. PCR-based amplicon sequencing targets specific genomic loci through primer amplification. Adaptive sampling differs from these approaches in several key aspects:

Adaptive sampling offers the advantage of minimal wet-lab overhead, making it particularly attractive for field-based or point-of-care applications where rapid turnaround is critical [2, 8]. However, its enrichment is modest compared to PCR, and it requires a high-quality reference index [12, 3].

Bioinformatics Workflow for Adaptive Sampling

The computational pipeline for adaptive sampling is integrated into the sequencing software and includes several modules. A simplified workflow is depicted below.

flowchart TD
    A[Raw ionic current signal], > B[Real-time basecalling]
    B, > C[Partial read alignment to reference index]
    C, > D{Match to target?}
    D, >|Yes| E[Continue sequencing full read]
    D, >|No| F[Reverse voltage, eject molecule]
    E, > G[Read output (FAST5/FASTQ)]
    F, > H[Reject read, record event]
    G, > I[Downstream analysis: assembly, variant calling, taxonomic classification]
    H, > A

Key bioinformatics considerations include:

• Reference index construction: The index must be computationally efficient for real-time use. Minimizer-based indexing reduces memory footprint and alignment time [6, 11]. Inclusion of decoy sequences (e.g., host genome, common contaminants) can improve specificity [12].

• Read length thresholds: Adaptive sampling decisions are made after a minimal number of bases are sequenced (typically 200-500 bp). Very short fragments may not yield reliable alignments and are often sequenced fully regardless of identity [10].

• Real-time resource management: The software must balance alignment speed with sequencing throughput. High target density or large reference indices can delay decisions and reduce pore utilization [12, 3].

• Post-sequencing analysis: Enriched reads can be subjected to metagenomic classification, de novo assembly, or targeted variant calling. The enrichment bias must be accounted for in quantitative analyses [11, 15].

Performance in Veterinary Pathogen Detection

Tick-Borne Pathogens

Adaptive sampling has been evaluated for detection of Borrelia burgdorferi in blacklegged ticks (Ixodes scapularis). Cassens et al. [1, 16] demonstrated that nanopore sequencing with adaptive sampling could detect B. burgdorferi DNA in tick lysates with sensitivity comparable to qPCR. Enrichment of spirochete sequences against a high background of tick and environmental DNA was achieved with a 10-15 fold increase in on-target reads compared to standard metagenomic sequencing [1, 16]. The method allowed strain-level genotyping through long-read assembly of the ospC and 16S-23S rRNA intergenic spacer regions.

In a related study, Kipp et al. [10] applied adaptive sampling to sequence mitochondrial genomes of hematophagous insects and simultaneously identify bloodmeal sources. By targeting both insect mitochondrial barcodes and vertebrate hemoglobin genes, the authors achieved dual enrichment that enabled vector-host interaction studies without separate library preparation [10].

Zoonotic and Carnivore Pathogens

Elshafie et al. [2] developed a targeted sequencing panel for carnivore pathogens, including SARS-CoV-2, canine distemper virus, and feline calicivirus. The panel employed adaptive sampling with a reference index containing whole-genome sequences of 15 common carnivore viruses. In spiked samples and clinical specimens, adaptive sampling increased viral read proportions from <1% to 15-30% of total reads, enabling complete genome recovery from samples with Ct values up to 32 [2]. The method was validated on nasal swabs and fecal samples from dogs and cats.

Hawes et al. [14] evaluated probe-based enrichment paired with nanopore sequencing for zoonotic viruses (e.g., rabies, hantavirus). While their work focused on hybrid capture, they noted that adaptive sampling could serve as a complementary strategy for viruses with high sequence diversity, provided the reference index captures the relevant genetic variability [14].

Foodborne and Livestock Pathogens

Buytaers et al. [11] demonstrated adaptive sampling for strain-level characterization of Salmonella enterica and Escherichia coli in wheat flour without culture enrichment. By targeting specific serotype marker genes and virulence loci, they achieved sufficient coverage to distinguish closely related strains and infer antimicrobial resistance gene profiles [11]. This application has direct relevance to veterinary food safety investigations.

Herbert et al. [3] examined the impact of microbiological methodologies on adaptive sampling performance in metagenomic studies. They found that sample preprocessing steps, such as host DNA depletion or microbial enrichment, significantly increased the efficiency of adaptive sampling by reducing the background read rejection burden [3]. For complex matrices like bovine feces, combined host depletion and adaptive sampling improved target detection by over 20-fold compared to sequencing alone.

Avian and Aquatic Applications

Adaptive sampling has been adapted for avian influenza virus surveillance. Although not explicitly covered in the provided literature, the principles described by Lin et al. [17, 8] for Chlamydia psittaci detection in birds can be extrapolated: real-time enrichment of pathogen reads from respiratory samples enabled rapid diagnosis and genomic surveillance. The approach has been integrated into workflows for Nanopore Sequencing for Real-Time Genomic Surveillance of Avian Influenza Viruses in Poultry (see /knowledge/diagnostics/nanopore-sequencing-avian-influenza-poultry-surveillance).

For aquatic systems, Kapoor et al. [4] adapted custom capture sequencing panels to the nanopore platform, demonstrating that long-read sequencing with targeted enrichment could identify viral pathogens in fish tissues. Although their study used hybrid capture, the authors noted that adaptive sampling could replace the capture step in future iterations to simplify the protocol.

Limitations and Challenges

Despite its advantages, adaptive sampling has several limitations [12, 3, 13].

• Modest enrichment factor: Typical enrichment ranges from 5 to 50-fold, which may be insufficient for samples with extremely low pathogen loads (<100 genome copies per reaction). In such cases, PCR-based approaches remain more sensitive [12, 13].

• Dependence on reference quality: The reference index must accurately represent the target sequence space. For highly variable RNA viruses or novel pathogen variants, mismatches can lead to false rejections and reduced sensitivity [2, 14].

• High host DNA burden: In samples with abundant host nucleic acid (e.g., tissue biopsies, blood), the proportion of rejected reads can be very high, reducing overall throughput and prolonging sequencing time [3, 15]. Host depletion prior to library preparation is often recommended [3].

• Read length bias: Adaptive sampling decisions rely on early read segments. Very long reads may be rejected before full sequencing, while very short reads may not provide enough information for accurate alignment [10].

• Computational overhead: Real-time alignment requires significant CPU resources, and on some portable devices, this can limit the number of pores that are simultaneously active [6, 12].

Future Directions

Several developments are expected to enhance adaptive sampling for veterinary diagnostics. Genome-guided generative adversarial learning, as described by Zhang et al. [6], can optimize reference indexes for complex or poorly characterized pathogens. The use of carrier DNA (NASCarD) has been shown to improve sequencing throughput by maintaining pore occupancy during adaptive rejections [13]. Integration with rapid PCR-based approaches, as demonstrated by Lin et al. [8] and Terrazos Miani et al. [13], can combine the sensitivity of amplification with the flexibility of adaptive sequencing. Finally, the development of pan-pathogen reference panels for common veterinary syndromes (e.g., bovine respiratory disease complex, canine infectious respiratory disease) would allow adaptive sampling to serve as a broad diagnostic screen [2, 15].

Frequently Asked Questions

How does nanopore adaptive sampling differ from traditional capture-based enrichment?

Adaptive sampling uses real-time bioinformatic selection to retain or eject individual DNA molecules as they pass through the nanopore, requiring no physical capture, washing, or amplification steps [6, 8]. Capture-based methods require probe hybridization and bead purification, adding several hours to the workflow [4, 14].

Can adaptive sampling detect novel or highly divergent pathogens?

The reference index must contain sequences similar to the target. For highly divergent pathogens, mismatches may cause reads to be rejected. Pan-family or genus-level reference sequences can mitigate this risk [2, 10]. Generative adversarial network-based index design may further improve detection of unknowns [6].

What sample types are most suitable for adaptive sampling in veterinary diagnostics?

Samples with moderate to high pathogen loads relative to host background (e.g., tick lysates, fecal samples, nasal swabs) perform well [1, 2, 11]. Tissues with high host DNA content (e.g., spleen, lymph node) benefit from pre-sequencing host DNA depletion [3, 15].

Is adaptive sampling quantitative?

Enrichment by adaptive sampling is not linear across abundance levels, so direct quantification of pathogen load is not reliable without spiked internal standards [12, 11]. Relative comparisons between samples using normalized coverage can be cautiously interpreted.

What bioinformatics expertise is required to implement adaptive sampling?

Basic familiarity with command-line tools for reference index construction and real-time sequence data management is needed. Publicly available pipelines and cloud-based solutions are lowering the barrier to entry [6, 11].

References

[1] Cassens J, Kipp EJ, Frank LE, et al. Evaluating the detection capabilities of nanopore sequencing for Borrelia burgdorferi detection in blacklegged ticks. Sci Rep. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/41807552/

[2] Elshafie NO, Kattoor JJ, Kelly J, et al. MinION Adapted tNGS Panel for Carnivore Pathogens Including SARS-CoV-2. Pathogens. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41599007/

[3] Herbert J, Thompson S, Beckett AH, et al. Impact of microbiological molecular methodologies on adaptive sampling using nanopore sequencing in metagenomic studies. Environ Microbiome. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40325409/

[4] Kapoor V, Sanchez-Vicente S, Donovan W, et al. Adaptation of custom capture sequencing panels to the Oxford Nanopore MinION platform. Mol Biol Rep. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/41774281/

[5] Eckert SE, Chan JZ, Houniet D, et al. Enrichment by hybridisation of long DNA fragments for Nanopore sequencing. Microb Genom. 2016. URL: https://pubmed.ncbi.nlm.nih.gov/28785419/

[6] Zhang Y, Sun P, Zhang J, et al. Genome-guided generative adversarial learning enables nanopore adaptive sequencing. Nat Commun. 2026. URL: https://pubmed.ncbi.nlm.nih.gov/42230586/

[7] Filser M, Schwartz M, Merchadou K, et al. Adaptive nanopore sequencing to determine pathogenicity of BRCA1 exonic duplication. J Med Genet. 2023. URL: https://pubmed.ncbi.nlm.nih.gov/37263769/

[8] Lin Y, Dai Y, Liu Y, et al. Rapid PCR-Based Nanopore Adaptive Sequencing Improves Sensitivity and Timeliness of Viral Clinical Detection and Genome Surveillance. Front Microbiol. 2022. URL: https://pubmed.ncbi.nlm.nih.gov/35783376/

[9] Quick J, Loman NJ, Duraffour S, et al. Real-time, portable genome sequencing for Ebola surveillance. Nature. 2016. URL: https://pubmed.ncbi.nlm.nih.gov/26840485/ *** Disclaimer: This article is for educational and informational purposes only. It is not intended to substitute for professional veterinary advice, diagnosis, treatment, or regulatory guidance. Always consult a licensed veterinarian or qualified specialist regarding animal health, disease diagnosis, and therapeutic decisions.

[10] Kipp EJ, Lindsey LL, Milstein MS, et al. Nanopore adaptive sampling for targeted mitochondrial genome sequencing and bloodmeal identification in hematophagous insects. Parasit Vectors. 2023. URL: https://pubmed.ncbi.nlm.nih.gov/36788607/

[11] Buytaers FE, Verhaegen B, Van Nieuwenhuysen T, et al. Strain-level characterization of foodborne pathogens without culture enrichment for outbreak investigation using shotgun metagenomics facilitated with nanopore adaptive sampling. Front Microbiol. 2024. URL: https://pubmed.ncbi.nlm.nih.gov/38495515/

[12] Xu W, Top J, Viveen MC, et al. Limited value of Nanopore adaptive sampling in a long-read metagenomic profiling workflow of clinical sputum samples. BMC Med Genomics. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41327286/

[13] Terrazos Miani MA, Borcard L, Gempeler S, et al. NASCarD (Nanopore Adaptive Sampling with Carrier DNA): A Rapid, PCR-Free Method for SARS-CoV-2 Whole-Genome Sequencing in Clinical Samples. Pathogens. 2024. URL: https://pubmed.ncbi.nlm.nih.gov/38251368/

[14] Hawes K, Greene B, Weishampel ZA, et al. Evaluation of a Probe-Based Enrichment Protocol for Nanopore Sequencing of Zoonotic Viruses. Viruses. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/41305486/

[15] Ring N, Low AS, Wee B, et al. Rapid metagenomic sequencing for diagnosis and antimicrobial sensitivity prediction of canine bacterial infections. Microb Genom. 2023. URL: https://pubmed.ncbi.nlm.nih.gov/37471128/

[16] Cassens J, Kipp EJ, Frank LE, et al. Evaluating the diagnostic capabilities of nanopore sequencing for Borrelia burgdorferi detection in blacklegged ticks. bioRxiv. 2025. URL: https://pubmed.ncbi.nlm.nih.gov/40909603/

[17] Lin Y, Dai Y, Zhang S, et al. Application of nanopore adaptive sequencing in pathogen detection of a patient with Chlamydia psittaci infection. Front Cell Infect Microbiol. 2023. URL: https://pubmed.ncbi.nlm.nih.gov/36756615/

[18] Forghani F, Li S, Zhang S, et al. Salmonella enterica and Escherichia coli in Wheat Flour: Detection and Serotyping by a Quasimetagenomic Approach Assisted by Magnetic Capture, Multiple-Displacement Amplification, and Real-Time Sequencing. Appl Environ Microbiol. 2020. URL: https://pubmed.ncbi.nlm.nih.gov/32358002/