Sequencing has transformed our ability to understand infectious disease. Pathogen genomes can help distinguish lineages, investigate outbreaks, monitor antimicrobial resistance and reveal how infections move through populations. But whole-genome sequencing can require expensive equipment, specialised laboratory processes, significant computational infrastructure and technical expertise. For pathogens that cannot easily be cultured, the challenge can be greater still.
In July 2026, researchers from the Wellcome Sanger Institute and collaborators in Zimbabwe and South Africa published an interesting attempt to change this equation. Their approach, called Phylo-Plex, asks a deceptively simple question: what if genomic surveillance did not always require sequencing the whole genome?
The researchers demonstrated a method capable of extracting much of the epidemiologically useful information needed to distinguish pathogen lineages while sequencing only a small fraction of the genome. For Treponema pallidum, the bacterium responsible for syphilis, they demonstrated the approach in a laboratory in Harare, Zimbabwe. Their estimated laboratory cost: £12.47 per sample. That creates another interesting question — what happens to genomic surveillance infrastructure when sequencing becomes cheap enough to deploy almost anywhere?
Whole-genome sequencing is enormously powerful because it captures information across the pathogen genome. But not every surveillance question necessarily requires every nucleotide. Suppose we want to determine which lineage of a pathogen is circulating in a population — some parts of its genome may provide considerably more information about that lineage than others.
Phylo-Plex uses existing whole-genome data to identify those particularly informative regions. The process begins with a representative collection of pathogen genomes. Researchers identify single-nucleotide polymorphisms — individual positions in the genome where sequences differ — that distinguish branches of the pathogen’s phylogenetic tree. Those discriminatory positions are then grouped geographically along the genome so that a relatively small number of regions captures as much phylogenetic information as possible. Primers can then be designed to amplify those regions.
Instead of sample → entire pathogen genome → analysis, the surveillance system can become sample → selected information-rich genomic regions → lineage reconstruction. The objective is not to replace whole-genome sequencing universally. It is to obtain enough genomic information to answer particular surveillance questions much more economically.
The researchers tested the concept particularly extensively using Treponema pallidum. Their final scheme contained 59 multiplex amplicons which, together, covered only approximately 3.6 per cent of the genome. Yet the researchers found that this selected fraction could reproduce the broad population structure obtained using whole-genome information with a high degree of concordance. In laboratory testing, variant calls generated from Oxford Nanopore sequencing of the selected regions were also highly correlated with variants obtained using Illumina whole-genome sequencing.
This is an important distinction. Phylo-Plex is not claiming that 3.6 per cent of a genome contains everything that could be learned from the whole genome — it does not. Instead, the method deliberately selects genomic regions that contain disproportionately useful information for distinguishing pathogen lineages. It is genomic surveillance designed around the question being asked.
The choice of pathogen is important. T. pallidum is difficult to culture routinely in the laboratory, which means sequencing may need to be performed directly from clinical material. But in a clinical swab, pathogen DNA can represent a tiny fraction of the total DNA present — the Sanger team notes that in some specimens it may account for as little as 0.01 per cent of total DNA. Recovering a complete pathogen genome under those circumstances can require enrichment or metagenomic approaches that add cost and complexity.
Targeting carefully selected regions changes the problem. Rather than attempting to recover everything, the assay concentrates on the information required for high-resolution lineage tracking — making the method particularly interesting for laboratories where the conventional route to whole-genome sequencing may be difficult to sustain at scale.
The headline cost deserves careful treatment. The researchers estimated their 59-amplicon T. pallidum Phylo-Plex protocol at £12.47 per sample when samples were processed in batches of 24. That figure included the reagents and consumables required for sequencing PCR-positive DNA extracts. It did not include everything required to establish a genomic surveillance laboratory: the paper explicitly excludes one-time setup costs such as purchasing the sequencer and a suitable high-performance laptop, and DNA extraction and quality-control costs are additional. The costing was also calculated using publicly available UK list prices, meaning actual costs in different countries may vary because of procurement, shipping, importation and local supply chains.
Those qualifications matter. £12.47 should not be interpreted as “you can establish genomic surveillance for £12.47.” It means that once the necessary infrastructure and processes exist, the estimated consumable cost for this particular assay and batch configuration was £12.47 per PCR-positive sample. That is still significant: for comparison, the researchers estimated conventional commercial Sanger-based multilocus sequence typing at approximately £28–£112 per sample depending on reaction cost, and note that commercial bacterial whole-genome sequencing in the UK can be around £65 per sample, often relying on large batches to achieve economical pricing. Phylo-Plex can operate in batches of 24 for less than £300 in the researchers’ costing model, with the workflow returning results within approximately 24–48 hours.
Cost per sample is only part of the economics of surveillance. Batch size matters too. Large sequencing platforms can achieve impressive per-sample economics when hundreds or thousands of samples are processed together, but surveillance does not always generate samples at convenient intervals. A laboratory may have twelve samples today; waiting several weeks until there are enough samples to fill a sequencing run may reduce costs, but it may also reduce the usefulness of the information. During an emerging outbreak, the value of a result can decline rapidly with time.
A method capable of operating economically with smaller batches therefore changes more than the laboratory budget — it potentially changes the latency of surveillance. Instead of collect → store → accumulate batch → sequence → analyse, the system can move closer to collect → sequence → analyse → respond. That is an important architectural difference.
Phylo-Plex was demonstrated using Oxford Nanopore MinION technology with Flongle flow cells. Portable nanopore sequencing has already changed assumptions about where sequencing needs to happen — a sequencing instrument no longer necessarily means a room-sized or even bench-sized machine located inside a major genomics centre.
But making the sequencer portable solves only one component of the problem. A functioning surveillance system still requires samples, reagents, laboratory processes, quality control, trained people, computing, reference data, analytical workflows, storage, connectivity, metadata, and mechanisms for communicating results. This is where the next phase of the genomic-surveillance challenge begins.
Reducing sequencing costs is unquestionably valuable. But success creates its own infrastructure problem. Imagine genomic surveillance becoming sufficiently inexpensive and technically accessible that hundreds of regional laboratories can participate rather than a small number of national centres. Instead of genomic information flowing from 10 laboratories, we might eventually have it flowing from 100, or 1,000.
That is a tremendous scientific opportunity. It also creates a data-management problem. Who identifies the samples consistently? How are specimen and patient or host metadata represented? How are results compared between laboratories? Which assay version generated a sequence, which primer scheme, which sequencing chemistry, which basecaller, which analytical pipeline, which reference genome, which software versions, which quality thresholds? When was the analysis performed, who changed a record, which data can be shared, and can somebody reproduce the result later? Reducing the cost of generating genomic information therefore increases the importance of the infrastructure surrounding it.
A central sequencing facility has some inherent advantages: processes can be standardised, equipment can be controlled, analytical pipelines can be centrally managed, and quality systems can be applied consistently. A distributed genomic network is different. Laboratories may use different equipment, different staff may process samples, software may be updated at different times, connectivity may vary, local data-governance requirements may differ, reference datasets may change, and some facilities may be offline for periods of time.
The objective cannot therefore simply be to make every laboratory identical. Instead, a distributed surveillance system needs mechanisms that allow independently generated information to remain interpretable and trustworthy. That requires provenance. For every result we may need to establish: where the biological sample originated, how it was processed, which assay generated the data, which instrument produced the sequence, which computational workflow analysed it, which reference information was used, what quality criteria were applied, which result was produced, and how that result entered the surveillance system. The more distributed sequencing becomes, the more valuable that chain becomes — the same principle behind how we think about <a href="”>connecting provenance across systems at The BioChain, whatever the pathogen or platform.
Phylo-Plex demonstrates another important principle: its targeted regions are selected using existing pathogen genome collections. In other words, cheap targeted surveillance depends partly upon the genomic knowledge generated previously through whole-genome sequencing. This creates a productive relationship between the two approaches. Whole-genome sequencing can continue to discover diversity, identify emerging lineages and improve our understanding of pathogen populations; that knowledge can then inform targeted surveillance systems designed for scale.
The researchers themselves describe a model in which a subset of samples undergoes whole-genome sequencing to identify important lineages, followed by Phylo-Plex deployment for larger-scale tracking — a surveillance cycle running from whole-genome sequencing to population understanding, informative regions, targeted surveillance, detection of a new or unusual pattern, selected samples returned to whole-genome sequencing, and an updated population model. That is considerably more sophisticated than simply choosing between “cheap sequencing” and “expensive sequencing.” They can become complementary layers of the same surveillance architecture.
Pathogens evolve. A genomic region that is highly informative today may not necessarily remain sufficient indefinitely: new lineages emerge, population structures change, previously rare variants become common, new resistance mutations appear, and geographic expansion introduces populations that were poorly represented in the original reference dataset. Phylo-Plex was deliberately designed with this problem in mind — existing schemes can be modified, and additional pools of amplicons can potentially be introduced to address new outbreaks or lineages.
But that creates a versioning problem. Imagine Laboratory A uses Phylo-Plex Scheme 1.0, Laboratory B uses Scheme 1.2, and Laboratory C has added an outbreak-specific amplicon pool. Their results may all be scientifically valid, but the surveillance platform needs to know that they are not necessarily identical observations. The assay definition itself therefore becomes metadata. Versioning becomes part of epidemiology.
The Phylo-Plex researchers have made their computational methods available publicly, and their published workflow evolved during development. Initial analysis used a collection of tools including Cutadapt, Clair3 and bcftools; the researchers subsequently implemented a streamlined Nextflow pipeline, incorporating Dorado basecalling, additional quality-control steps and refinements to variant calling and pseudosequence generation.
That is good scientific practice. It also illustrates why genomic provenance cannot end with the sequencing instrument. The same raw reads can pass through different computational workflows — software evolves, models change, quality procedures improve. The final genomic result therefore needs to remain connected to the computational process that created it. The laboratory protocol and computational workflow are both part of the evidence chain.
The Phylo-Plex study provides a strong example of reproducibility infrastructure. Sequencing data have been deposited in the European Nucleotide Archive, with accession identifiers provided for the underlying datasets. Code used to identify discriminatory sites and perform the hierarchical selection process is publicly available, the analytical workflow is available as a Nextflow pipeline, and the code associated with the paper has also been archived through Zenodo.
These practices matter because they allow other researchers to inspect, reproduce and extend the work. They also demonstrate something fundamental about modern genomics: a scientific publication is increasingly only the visible end of a much larger digital research object. Behind the paper sit samples, sequence datasets, metadata, software, workflows, reference information, versions, parameters, results, and the relationships connecting them. Preserving those relationships is becoming as important as preserving the individual files.
For years, a central question in genomics has been: how can we make sequencing cheaper? That remains important. But technologies such as Phylo-Plex suggest that we should begin asking the next question — what happens when we succeed? If genomic surveillance becomes cheap enough to deploy across hundreds or thousands of laboratories, the bottleneck begins moving: towards interoperability, towards metadata, towards quality assurance, towards computational reproducibility, towards data governance, towards secure sharing, towards long-term storage, and towards provenance. The problem changes from “can we generate genomic information?” to “can we create a trustworthy network from all the genomic information we are now capable of generating?”
These are not the same thing. A country could theoretically possess dozens of sequencing instruments without possessing an effective genomic-surveillance system. Surveillance capacity requires the ability to turn biological samples into timely, comparable and trustworthy evidence — connecting laboratory capacity with data infrastructure across the full chain: sample → assay → sequence → quality control → analysis → provenance → interpretation → sharing → surveillance → decision.
Every stage matters. If sequencing becomes dramatically cheaper while everything after sequencing remains fragmented, we risk producing enormous quantities of genomic data without creating an equivalent increase in actionable knowledge.
Phylo-Plex is not a universal replacement for whole-genome sequencing, nor do its authors present it as one. Its importance lies somewhere more interesting: it demonstrates that high-resolution pathogen surveillance can potentially be designed around the minimum genomic information required to answer a particular epidemiological question. For T. pallidum, the researchers recovered only a small fraction of the genome while retaining substantial lineage-discrimination capability, demonstrated the workflow in Zimbabwe, achieved a laboratory consumables estimate of £12.47 per PCR-positive sample under their stated assumptions, and produced results within one to two days. Those are meaningful steps towards making genomic epidemiology more deployable.
But if approaches like this succeed, genomic surveillance will become increasingly distributed, and distributed genomic surveillance needs more than distributed sequencers. It needs distributed infrastructure for trust. We need to know where a sample came from, which method produced its sequence, which version of that method was used, how the sequence was analysed, and whether two results generated thousands of kilometres apart are genuinely comparable. And years later, we need to be able to reconstruct how a surveillance conclusion was reached.
The future of genomic surveillance may therefore depend upon solving two problems simultaneously: making genomic information cheap enough to generate everywhere, and making genomic evidence trustworthy enough to use everywhere. Phylo-Plex takes an important step towards solving the first. The second may prove just as important.
Phylo-Plex is a low-cost amplicon sequencing method, developed by the Wellcome Sanger Institute and collaborators in Zimbabwe and South Africa, that targets only the most phylogenetically informative regions of a pathogen genome instead of sequencing the whole genome.
No. That figure covers only the reagents and consumables for the 59-amplicon T. pallidum protocol run in batches of 24, based on UK list prices. It excludes the sequencer, a suitable laptop, DNA extraction and quality-control costs, and actual costs will vary by country.
The method isn’t claiming that 3.6% of the genome contains everything the whole genome does. It deliberately selects the regions that carry disproportionate value for distinguishing pathogen lineages, and the researchers found this reproduced whole-genome population structure with high concordance for their test case.
It’s difficult to culture in the lab, so sequencing often has to be done directly from clinical material, where pathogen DNA can be as little as 0.01% of total DNA in a swab — making full genome recovery costly and complex, and a targeted approach comparatively more valuable.
If genomic surveillance becomes cheap enough for hundreds or thousands of laboratories to participate, results now need to be comparable and trustworthy across all of them — consistent sample identification, metadata, assay versions, pipelines and quality thresholds become the bottleneck instead of sequencing cost.
Their results can all be scientifically valid but not directly comparable unless the specific scheme version (and any added amplicon pools) is recorded and travels with the result as metadata — the assay version itself becomes part of the evidence.
No, and the researchers don’t present it that way. They describe a complementary model where whole-genome sequencing continues to identify emerging lineages and update the reference dataset, while Phylo-Plex handles larger-scale, cheaper ongoing tracking informed by that reference data.
If your organisation generates, manages, analyses or governs biological evidence and would be interested in participating in a UK or European provenance demonstrator, The BioChain would welcome the conversation.
Get in touch
Regulatory developments, technical notes and platform news — sent occasionally, straight to your inbox.