tNGS vs mNGS vs WGS for Pathogen Projects: Which Workflow Fits Your Study?

Figure 1: Three NGS approaches for pathogen research, distinguished by breadth of detection and depth of characterization.

You have a pathogen-focused project. Maybe you are tracking antimicrobial resistance genes across a hospital cohort, searching for the causative agent in culture-negative samples, or reconstructing transmission chains during an outbreak. You know next-generation sequencing can help. What is less obvious is which NGS approach — targeted (tNGS), metagenomic (mNGS), or whole genome sequencing (WGS) — matches your specific research question.

These three methods sit on a spectrum from narrow-and-deep to broad-and-shallow. Choosing wrong means overspending on coverage you do not need, or missing the pathogen you are looking for because your method was not designed to find it. This guide compares the three workflows side by side and provides a decision framework for matching method to research objective.

Three Technologies, Three Different Questions

Before comparing performance metrics, it helps to understand what question each method is built to answer.

Targeted NGS (tNGS) asks: "Which of the pathogens on my list are present, and do they carry resistance genes?" It uses multiplex PCR primers or hybridization probes to selectively amplify or capture nucleic acid from a predefined set of pathogens — typically dozens to a few thousand — directly from a clinical or environmental sample. Because it enriches only targets of interest, it requires far fewer sequencing reads and generates less background noise than untargeted approaches.

Metagenomic NGS (mNGS) asks: "What is in this sample — bacteria, fungi, viruses, parasites — whether I expect it or not?" It sequences all DNA (or RNA) in a sample without prior selection, producing millions of reads that are then computationally sorted into host and microbial fractions. This hypothesis-free design makes mNGS the method of choice when the pathogen is unknown, unexpected, or mixed.

Whole genome sequencing (WGS) asks: "What is the complete genetic makeup of this specific isolate?" It starts from a cultured or enriched pathogen, assembles its entire genome, and delivers the highest-resolution view available — every gene, every resistance determinant, every plasmid, and the single-nucleotide polymorphisms that distinguish one strain from another.

The three methods are complementary, not competitive. A well-designed research program may use all three: tNGS for screening, mNGS for discovery, and WGS for deep characterization.

Targeted NGS: When You Know What You Are Looking For

tNGS works by enriching pathogen-specific sequences before sequencing, which means the vast majority of reads map to targets rather than host background. This enrichment step is what gives tNGS its cost and speed advantages — and also its fundamental limitation.

Two enrichment strategies dominate. Multiplex PCR-based tNGS uses hundreds to thousands of primer pairs in a single reaction to amplify target regions from a panel of pathogens. It costs roughly $130 to $300 per sample and can return results in 12 to 14 hours, making it the fastest and least expensive NGS option for pathogen detection. Hybrid capture-based tNGS uses biotinylated DNA or RNA probes to physically pull down target sequences from a sequencing library. It covers a broader pathogen spectrum — up to several thousand species — and tolerates sequence mismatches better than PCR, but costs more and requires a longer hybridization step.

The strengths of tNGS are concrete. Because it reads target regions at high depth, it detects antimicrobial resistance (AMR) genes and resistance-conferring mutations more reliably than mNGS, which may lack sufficient coverage at those loci. A 2024 head-to-head comparison in lower respiratory tract infections found that capture-based tNGS identified 678 pathogenic strains versus 535 by mNGS, with sensitivity reaching 99.4% for panel targets. The method also handles DNA and RNA viruses in a single workflow, which is valuable for respiratory and CNS pathogen panels.

The limitation is equally concrete: tNGS cannot detect a pathogen that is not on the panel. If the causative agent is novel, unexpected, or missing from the probe set, tNGS returns a negative result regardless of what is actually in the sample. For this reason, clinical research protocols commonly specify that a negative tNGS result in a patient with strong evidence of infection should trigger reflex testing with mNGS.

Metagenomic NGS: Casting the Widest Net

mNGS sequences everything in a sample without selection — microbial DNA, host DNA, and environmental background alike. This comprehensiveness is its defining strength and its primary operational challenge.

In research settings, mNGS has demonstrated the ability to detect bacteria, fungi, DNA viruses, RNA viruses, and parasites from a single specimen, including organisms that standard culture misses entirely. A 2024 analysis of 520 tissue samples from suspected infections found that mNGS identified a pathogen in 85% of cases where conventional testing was negative, with clinical management changed in nearly half of all cases. In fever of unknown origin studies, mNGS has identified causative agents in roughly 29% more cases than culture-based methods alone.

The practical challenge is that in most sample types — particularly tissue, respiratory specimens, and body fluids — human DNA can account for over 90% of sequenced reads. This means the effective microbial sequencing depth is far lower than the nominal depth, and low-abundance pathogens can be masked. Two sample preparation strategies address this: whole-cell DNA (wcDNA) approaches that pellet and extract microbial cells, and cell-free DNA (cfDNA) approaches that sequence circulating microbial DNA from plasma. Each has trade-offs in sensitivity and concordance with culture, and the choice depends on sample type and research question.

The other well-documented challenge is specificity. mNGS detects everything present — including reagent contaminants, skin commensals introduced during collection, and latent viruses such as EBV or CMV that may be incidental rather than causal. Without careful negative controls, contamination tracking, and clinical contextualization, mNGS results can mislead. A 2024 review in the Journal of Clinical Investigation emphasized that accurate reference genome databases and rigorous contamination controls are essential prerequisites for interpretable mNGS data, noting that database errors propagate across all NGS methods that depend on them for taxonomic assignment.

Whole Genome Sequencing: Depth Over Breadth

If tNGS and mNGS are about detection, WGS is about characterization. Starting from a cultured isolate or an enriched sample, WGS assembles the complete genome of a single pathogen — chromosome, plasmids, and all — at single-nucleotide resolution.

This depth of information supports applications that neither tNGS nor mNGS can match. In antimicrobial resistance surveillance, WGS captures the full resistome: every acquired resistance gene, every chromosomal mutation, and the plasmid replicons that mobilize them between strains. Long-read platforms such as Oxford Nanopore and PacBio have become particularly valuable here, as they can resolve plasmid structures and gene duplications that short-read assemblies fragment. A 2025 study demonstrated a rapid ONT-based WGS pipeline that delivered actionable transmission data within two days at approximately $45 per isolate in materials cost.

For outbreak investigation, WGS provides the phylogenetic resolution needed to distinguish a true transmission cluster from genetically unrelated isolates that happen to share the same MLST type or serotype. This matters operationally: studies have shown that over 10% of isolates flagged as potentially related by conventional criteria belong to genetically distinct clusters, while roughly 44% of true genomic clusters lack the epidemiological links that trigger conventional investigation. Without WGS-level resolution, infection control resources are spent chasing pseudo-outbreaks while real transmission chains continue undetected.

The trade-off is that WGS requires a cultured isolate or a sufficiently enriched sample, which adds days to the workflow and limits its use to organisms that can be grown or captured. For unculturable pathogens or polymicrobial infections where the relevant organism cannot be isolated, tNGS or mNGS remain necessary.

Matching Method to Research Question

The table below maps common pathogen research objectives to the most appropriate sequencing approach. Use it as a starting point, not a rigid rule — many projects benefit from combining methods.

For projects that span multiple objectives, a tiered strategy often works best. For example, a respiratory pathogen research program might use tNGS with a broad respiratory panel as a first-pass screen on all enrolled samples, reserve mNGS for tNGS-negative cases where clinical suspicion remains high, and apply WGS to cultured isolates from confirmed cases for detailed AMR and phylogenetic analysis.

Cost, Turnaround, and What Drives Them

Budget and timeline constraints often determine which method is feasible for a given study. The table below summarizes representative ranges based on published studies and service provider data.

Several factors can shift these numbers substantially. For mNGS, samples with high host DNA content — tissue biopsies, respiratory specimens — may need deeper sequencing to achieve adequate microbial coverage, increasing cost. Adding host depletion or microbial enrichment steps adds $20 to $50 per sample but can improve effective microbial read depth by orders of magnitude. For tNGS, panel size and enrichment chemistry are the main cost drivers; a focused 20-pathogen respiratory panel costs less than a 3,000-target comprehensive panel. For WGS, the cost gap between short-read (Illumina) and long-read (ONT, PacBio) platforms has narrowed, but long-read sequencing still carries a premium for applications that require both high accuracy and long reads.

Turnaround time depends not only on the sequencing run itself but also on queue depth at the sequencing facility, bioinformatics pipeline runtime, and whether the project requires custom analysis beyond standard deliverables. A provider quoting "20-hour TAT" may be referring to sequencing time only, not the full sample-to-report interval. Confirming what the quoted turnaround includes — and excludes — prevents timeline surprises.

Scoping Your Pathogen Sequencing Project

A well-scoped project brief helps sequencing providers return comparable quotes and relevant feasibility feedback. Before reaching out, define these parameters:

Sample type and expected microbial load. Stool, respiratory specimens, tissue biopsies, CSF, plasma, and environmental samples each impose different extraction requirements, host depletion needs, and expected pathogen titers. A BAL fluid sample from a pneumonia patient is fundamentally different from a plasma sample from a sepsis workup, and the provider needs both the sample type and the expected organism classes to recommend appropriate protocols.

Research question and required resolution. Are you cataloging which pathogens are present (tNGS or mNGS), determining whether they carry specific resistance genes (tNGS or WGS), or tracing transmission at single-nucleotide resolution (WGS)? The answer determines not only the method but also the sequencing depth and bioinformatics scope.

Throughput and batch design. A project with 500 samples benefits from different economies of scale than one with 20. For mNGS and tNGS, larger batches reduce per-sample library preparation and sequencing costs through multiplexing. For WGS, per-isolate costs drop when samples are batched on high-throughput platforms. Discuss batch size and expected total sample count with the provider upfront — pricing often shifts at thresholds of roughly 50, 100, and 500 samples.

Bioinformatics deliverables. Standard deliverables for tNGS typically include a pathogen detection report with read counts and AMR gene calls where applicable. mNGS standard delivery should include taxonomic classification at species or genus level, estimated relative abundance, and negative control results. WGS deliverables should specify assembly statistics (N50, total length, estimated completeness), AMR gene annotation, and, for outbreak projects, a phylogenetic tree or distance matrix. If your team has in-house bioinformatics capacity, you may only need raw or pre-processed sequencing data; if not, confirm that the provider's standard pipeline produces analysis-ready outputs for your research question.

For projects that need end-to-end support — from nucleic acid extraction through annotated genome assemblies — a dedicated pathogen sequencing service that bundles wet-lab and bioinformatics workflows reduces the coordination burden and keeps accountability in one place.

Figure 2: Decision pathway for selecting a pathogen NGS approach based on research objective and sample context.

Figure 3: A tiered pathogen sequencing strategy that balances breadth, depth, and cost by sequencing most samples with tNGS while reserving mNGS and WGS for cases that need them.

FAQ

Can I use tNGS and mNGS on the same sample?

Yes. Many research protocols use tNGS as a first-line screen followed by mNGS on tNGS-negative samples where the investigator still suspects an infectious etiology. Because tNGS costs roughly one-quarter to one-half of mNGS, this tiered strategy reduces total project cost while preserving the ability to detect off-panel pathogens when needed. The two methods can also be run in parallel on split aliquots when both broad detection and targeted AMR profiling are required.

How does host DNA affect my mNGS results, and what can I do about it?

Human DNA routinely exceeds 90% of reads in tissue, respiratory, and body fluid samples, which means a 20-million-read mNGS run may yield fewer than 2 million microbial reads — and sometimes far fewer. Options to improve microbial yield include: selecting a host depletion kit (differential lysis, methylation-based depletion, or DNase treatment), using whole-cell DNA extraction rather than cell-free DNA when feasible, and increasing sequencing depth to compensate for the host fraction. The choice depends on sample type; discuss options with the provider during project scoping.

How many reads do I need for mNGS to reliably detect a pathogen?

There is no universal number — the required depth depends on sample type, expected pathogen load, and host DNA fraction. In practice, 20 million read pairs is a common starting point for respiratory and tissue samples, while 50 million or more may be needed for samples with very high host background or when searching for low-titer organisms. For sterile-site samples such as CSF, lower depth can be sufficient because pathogen nucleic acid constitutes a higher fraction of total DNA. The provider should be able to recommend a depth based on prior experience with similar sample types.

Does WGS require a cultured isolate, or can it work directly from a clinical sample?

Standard WGS requires a cultured isolate or a highly enriched sample to ensure sufficient pathogen DNA without host background. Direct-from-sample WGS is possible in principle — particularly with selective lysis or capture methods — but is not yet routine and carries a substantial risk of incomplete or contaminated assemblies. For unculturable organisms, metagenomic assembly from deep mNGS data can sometimes recover draft genomes, but this approach requires high sequencing depth and careful binning, and the results should be treated as exploratory rather than definitive.

How should I handle mixed infections where multiple pathogens are suspected?

mNGS is the strongest option for characterizing polymicrobial infections because it detects all DNA present without prior selection. tNGS can detect co-infections if both organisms are on the panel, but the relative abundance estimates may be skewed by differential amplification efficiency across targets. WGS from a mixed sample typically requires isolating each organism first — either by culture or by computational binning from metagenomic data. If polymicrobial infection is central to your research question, plan for mNGS as the primary detection method and WGS on individual isolates for detailed characterization.

What reference databases are used for taxonomic classification, and does the choice matter?

Yes, database choice significantly affects taxonomic assignment, as demonstrated by a 2023 comparison of five widely used 16S analysis pipelines. For mNGS, common reference databases include Kraken2 with the RefSeq or GTDB databases, MetaPhlAn with its curated marker gene set, and various commercial curated databases. For AMR gene detection, CARD, ResFinder, and AMRFinderPlus are widely used. The database version and its curation standard directly affect both sensitivity and specificity — a pathogen absent from the database cannot be detected regardless of its abundance in the sample. Ask the provider which databases and versions they use, and confirm they are updated regularly.

Related CD Genomics Services

• Pathogen Sequencing Solutions — End-to-end NGS services spanning tNGS, mNGS, and WGS for pathogen detection, characterization, and surveillance research.
• Pathogens mNGS Sequencing Service — Hypothesis-free metagenomic sequencing for comprehensive pathogen detection from complex samples.

References

1. Rodino KG, Simner PJ. Status check: next-generation sequencing for infectious-disease diagnostics. Journal of Clinical Investigation. 2024;134(4):e178003. doi:10.1172/JCI178003
2. Yin Y, Zhu P, Guo Y, Li Y, Chen H, Liu J, Sun L, Ma S, Hu C, Wang H. Enhancing lower respiratory tract infection diagnosis: implementation and clinical assessment of multiplex PCR-based and hybrid capture-based targeted next-generation sequencing. eBioMedicine. 2024;107:105307. doi:10.1016/j.ebiom.2024.105307
3. Yi L, Tan L, Long Q, Lyu X, Zeng H, Peng Y, Gu D, Liu H, Ge H, Yu Y, Li Z, Hu M. Comparative diagnostic performance of metagenomic and two targeted sequencing methods in lower respiratory infection. Scientific Reports. 2025;15:27365. doi:10.1038/s41598-025-11834-w
4. Sun N, Zhang J, Guo W, Cao J, Chen Y, Gao D, Xia X. Comparative analysis of metagenomic next-generation sequencing for pathogenic identification in clinical body fluid samples. BMC Microbiology. 2025;25(1):165. doi:10.1186/s12866-025-03887-8
5. Matsumura Y, Yamamoto M, Gomi R, Tsuchido Y, Shinohara K, Noguchi T, Nagao M. Integrating whole-genome sequencing into antimicrobial resistance surveillance: methodologies, challenges, and perspectives. Clinical Microbiology Reviews. 2025;38(4):e00140-22. doi:10.1128/cmr.00140-22
6. Hiergeist A, Ruelle J, Emler S, Gessner A. Reliability of species detection in 16S microbiome analysis: comparison of five widely used pipelines and recommendations for a more standardized approach. PLOS ONE. 2023;18(2):e0280870. doi:10.1371/journal.pone.0280870

For Research Use Only. Not for use in diagnostic or therapeutic procedures.