Choosing the Right Sequencing Platform for Shotgun Metagenomics: Illumina, Nanopore, and PacBio Compared

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-20

Key Takeaways

• Illumina platforms excel in high-accuracy, short-read (150-300 bp) sequencing, making them ideal for species-level taxonomic profiling and functional gene prediction in complex communities, particularly when deep coverage is achievable for detecting rare taxa and antimicrobial resistance genes. However, their short reads limit de novo assembly contiguity, hindering strain-level resolution and the reconstruction of complete genomes.

• Oxford Nanopore Technologies (ONT) offers long-read (average 10-30 kb, ultralong possible) sequencing with real-time analysis capabilities, enabling highly contiguous assemblies crucial for recovering complete microbial genomes and resolving repetitive regions. Its direct detection of base modifications is valuable for epigenetic studies, but the higher per-base error rate necessitates robust error correction strategies, and some applications require higher DNA input.

• PacBio platforms provide high-accuracy long reads (15-25 kb HiFi) through circular consensus sequencing, combining the assembly advantages of long reads with the accuracy of short reads. This makes them superior for complete microbial genome reconstruction and detailed strain-level analysis, though their higher cost per gigabase and longer turnaround times limit their application in large-scale or time-sensitive studies.

• Platform choice is dictated by research objectives and sample constraints, with Illumina favored for broad taxonomic and functional profiling in well-characterized communities, Nanopore for rapid, contiguous assembly and real-time insights (e.g., outbreak surveillance), and PacBio for high-fidelity genome reconstruction and strain resolution when budget permits.

• Hybrid approaches, combining short and long reads, leverage the strengths of both technologies to achieve highly contiguous and accurate assemblies, enabling comprehensive genomic analysis of complex microbial communities, though this necessitates more complex bioinformatics pipelines.

Shotgun metagenomics sequences all DNA fragments in a sample without prior target selection, providing both taxonomic and functional information about entire microbial communities. The platform choice determines read length, accuracy, throughput, and cost, each of which affects assembly contiguity, taxonomic resolution, and error profiles. This article compares Illumina, Oxford Nanopore, and PacBio platforms specifically for shotgun metagenomic projects, with emphasis on how each platform's characteristics align with different research objectives.

Scope and Decision Context

Researchers face a practical problem when designing shotgun metagenomic studies: the sequencing platform must match the project's biological questions, sample characteristics, and budget constraints. Whole-genome shotgun sequencing detects bacterial species, increases diversity detection, and improves gene prediction compared to 16S amplicon methods, but the platform choice introduces distinct trade-offs [<a href="#ref-1">1</a>]. Short-read platforms such as Illumina produce high-accuracy reads of 150 to 300 base pairs, while long-read platforms such as Oxford Nanopore and PacBio produce reads from thousands to hundreds of thousands of base pairs with different error profiles.

The decision framework presented here applies to human microbiome studies, environmental samples, clinical diagnostics, wastewater surveillance, and agricultural microbiology. Each application imposes different requirements for turnaround time, cost per sample, detection limits, and bioinformatics complexity. The comparison draws on published benchmarking studies and official documentation from sequencing data resources and analysis platforms.

At a Glance: Platform Comparison for Shotgun Metagenomics

The table reflects general platform characteristics as documented in published comparisons and official resource descriptions. Specific instrument models within each platform family vary in throughput and cost, so researchers should verify current specifications from manufacturer documentation and local sequencing facilities.

Core Principles of Shotgun Metagenomic Sequencing

What Shotgun Metagenomics Measures

Shotgun metagenomics sequences random DNA fragments from an entire microbial community, capturing bacteria, archaea, fungi, viruses, and functional genes in a single assay [<a href="#ref-2">2</a>]. Unlike 16S rRNA amplicon sequencing, which targets a single conserved gene, shotgun approaches provide a hypothesis-free view of community composition and metabolic potential [<a href="#ref-3">3</a>]. The approach requires no prior knowledge of which organisms are present, making it suitable for discovering novel taxa and unexpected pathogens.

The method produces two primary data types: taxonomic profiles that describe which organisms are present and functional profiles that describe which genes and pathways are encoded [<a href="#ref-2">2</a>]. Both profiles derive from the same sequencing data but require different analysis approaches. Taxonomic profiling typically uses read classification against reference databases, while functional profiling uses gene prediction and annotation. Assembly-based approaches reconstruct longer genomic fragments from individual organisms, enabling strain-level comparisons and the discovery of novel biosynthetic gene clusters.

Why Platform Choice Matters

The sequencing platform determines the fundamental properties of the data, including read length, accuracy, and throughput. These properties directly affect downstream analysis outcomes. Longer reads improve assembly contiguity because they span repetitive regions and resolve ambiguous overlaps. Higher accuracy reduces the error rate in taxonomic classification and gene prediction. Higher throughput enables deeper sampling of complex communities, which matters for detecting rare taxa and low-abundance functional genes.

A comparative study of human fecal microbiomes found that increased read length, either from longer reads or from assembled contigs, improved the accuracy of species detection [<a href="#ref-1">1</a>]. The same study demonstrated that whole-genome shotgun sequencing outperformed 16S amplicon sequencing for species detection, diversity measurement, and gene prediction [<a href="#ref-1">1</a>]. These findings establish the biological rationale for choosing shotgun approaches and for considering read length as a key parameter.

The Role of Sequencing Depth

Sequencing depth, measured as the total number of bases sequenced per sample, determines the sensitivity of the assay. Deep sequencing provides more coverage of low-abundance organisms and enables detection of rare functional genes. Shallow sequencing at approximately 1 Gb per sample can provide taxonomic assignments at genus and species levels that closely resemble deep sequencing results, while remaining more cost-effective for large cohort studies [<a href="#ref-4">4</a>].

The choice of depth depends on the research question. Community profiling studies that focus on dominant taxa may require only shallow sequencing. Functional profiling studies that aim to reconstruct metabolic pathways or detect low-abundance resistance genes require deeper sequencing. Longitudinal studies with hundreds or thousands of samples may need to balance depth against cost, and shallow whole-metagenome shotgun sequencing offers a practical compromise [<a href="#ref-4">4</a>].

Illumina Platforms for Shotgun Metagenomics

Instrument Family and Output Characteristics

Illumina sequencing uses sequencing-by-synthesis chemistry to generate short reads with high base-level accuracy. The instrument family includes the MiSeq, HiSeq, NextSeq, and NovaSeq systems, each with different throughput and read length capabilities. The NextSeq 500 in high-output mode with 2x150 base pair paired-end sequencing has been used successfully for shotgun metagenomic diagnosis of corneal infections [<a href="#ref-3">3</a>]. The MiSeq and HiSeq platforms were compared directly in a human fecal microbiome study that established the reproducibility of shotgun methods across platforms [<a href="#ref-1">1</a>].

Illumina platforms dominate shotgun metagenomics because of their established accuracy, mature bioinformatics support, and wide availability in core facilities. The short-read data format is compatible with most metagenomic analysis tools, including those available through Bioconductor packages and Galaxy workflows [<a href="#ref-5">5</a>][<a href="#ref-6">6</a>]. The high accuracy of Illumina reads reduces the need for error correction and simplifies taxonomic classification.

Strengths for Metagenomic Applications

Short-read Illumina sequencing excels at species-level taxonomic profiling and functional gene prediction. The high accuracy enables confident assignment of reads to reference genomes and accurate detection of single nucleotide variants. Deep sequencing on high-throughput instruments such as the NovaSeq allows comprehensive sampling of complex communities, including rare taxa that would be missed at lower depths.

Illumina data are well suited for read-based analysis approaches that classify individual reads against reference databases. These approaches work well for well-characterized communities where reference genomes exist for most members. The high accuracy also supports sensitive detection of antimicrobial resistance genes and virulence factors, which is important for clinical applications [<a href="#ref-7">7</a>].

Limitations and Failure Modes

Short reads create challenges for de novo assembly of metagenomes. Repetitive regions, mobile genetic elements, and closely related strains produce ambiguous assemblies that fragment into many contigs. The assembly problem worsens in complex communities with many closely related species, where shared sequence regions prevent unambiguous reconstruction [<a href="#ref-8">8</a>].

Short reads also limit strain-level resolution. Distinguishing closely related strains requires variants that may fall outside individual read lengths. While paired-end reads provide some linkage information, the limited insert size restricts the ability to phase variants across longer genomic regions. Researchers studying strain dynamics or tracking specific isolates may need long-read data to resolve these questions.

Oxford Nanopore Platforms for Shotgun Metagenomics

Technology and Output Characteristics

Oxford Nanopore Technologies (ONT) sequencing passes DNA molecules through protein nanopores and measures changes in electrical current as bases pass through the pore. The technology produces long reads in real time, with the MinION providing portable sequencing and the PromethION providing higher throughput. The R10.4.1 flow cell chemistry has narrowed the accuracy gap with Illumina, though systematic differences remain [<a href="#ref-8">8</a>].

Nanopore sequencing offers unique advantages for metagenomics, including the ability to sequence very long DNA fragments. Ultralong reads can span entire mobile genetic elements, resolve repetitive regions, and produce highly contiguous assemblies. The real-time data generation enables adaptive sampling, where the instrument can selectively enrich or deplete specific taxa during the run.

Strengths for Metagenomic Applications

Long reads from Nanopore sequencing substantially improve assembly contiguity compared to short-read approaches. Complete or near-complete microbial genomes can be recovered from complex communities, enabling strain-level comparisons and the discovery of novel genomic elements. The technology has been used successfully for wastewater surveillance, where it enabled strain-level characterization of viruses including Influenza A and SARS-CoV-2 [<a href="#ref-9">9</a>].

Nanopore sequencing also detects base modifications directly, providing epigenetic information that is invisible to short-read platforms. This capability is valuable for studying gene regulation in microbial communities and for understanding how environmental conditions affect methylation patterns.

Limitations and Failure Modes

The per-base error rate of Nanopore sequencing remains higher than Illumina, despite improvements with newer chemistry. The R10.4.1 flow cell has narrowed but not eliminated the accuracy gap [<a href="#ref-8">8</a>]. Higher error rates complicate taxonomic classification, particularly for closely related species, and require error correction strategies during assembly.

Nanopore sequencing requires relatively high DNA input for optimal performance. Low-biomass samples such as aerosols, sputum, or dust present challenges because they yield limited DNA [<a href="#ref-10">10</a>][<a href="#ref-11">11</a>]. The technology also requires careful library preparation to avoid shearing long DNA molecules, which would negate the advantage of long reads.

PacBio Platforms for Shotgun Metagenomics

Technology and Output Characteristics

PacBio sequencing uses single-molecule real-time (SMRT) technology to observe DNA synthesis in real time. The HiFi mode produces circular consensus sequences by reading the same molecule multiple times, generating long reads with high accuracy. The Sequel and Revio instruments provide different throughput levels, with the Revio offering higher output.

PacBio HiFi reads combine the advantages of long read length with high accuracy, making them valuable for metagenomic assembly and strain resolution. The technology produces reads of 15 to 25 kilobases with accuracy comparable to short-read platforms. This combination enables complete microbial genome reconstruction from complex communities.

Strengths for Metagenomic Applications

HiFi reads produce highly contiguous assemblies with fewer errors than Nanopore-only approaches. The high accuracy reduces the need for polishing and simplifies downstream analysis. Complete or near-complete genomes recovered from metagenomes enable detailed comparisons of gene content, mobile elements, and structural variants.

PacBio sequencing also detects base modifications through the kinetics of DNA synthesis, providing epigenetic information similar to Nanopore. The technology is well suited for projects that require both long reads and high accuracy, such as characterizing novel species or tracking strain evolution.

Limitations and Failure Modes

PacBio sequencing costs more per gigabase than Illumina or Nanopore, limiting its use in large-scale studies. The higher cost makes deep sequencing of many samples prohibitive for most research groups. The technology also requires higher DNA input than short-read platforms, which can be problematic for low-biomass samples.

The longer turnaround time for PacBio sequencing may be a limitation for clinical applications that require rapid results. While Nanopore can provide real-time data during the run, PacBio requires the full sequencing run to complete before data analysis can begin.

Practical Workflow Considerations

Sample Preparation and DNA Extraction

DNA extraction method introduces variability that can affect downstream results. A systematic evaluation of four extraction methods found that DNA yield varied by method, with phenol:chloroform and Promega approaches producing the highest yields [<a href="#ref-10">10</a>]. The same study found that extraction method affected microbial community structure and functional profiles, though the impact varied by sample type [<a href="#ref-10">10</a>].

For low-biomass samples, extraction method choice becomes critical. A custom multi-component DNA isolation method optimized for aerosol samples improved DNA yields from filter-collected air samples by isolating DNA from the entire filter extract [<a href="#ref-11">11</a>]. The method combined chemical, enzymatic, and mechanical lysis to ensure comprehensive microbiome representation [<a href="#ref-11">11</a>].

Researchers should validate their extraction method using mock community controls to assess bias. The choice of extraction method should be documented and kept consistent across all samples in a study to minimize technical variation.

Library Preparation and Sequencing Depth

Library preparation protocols differ by platform and sample type. The NEBNext DNA ultra II protocol has been used successfully for shotgun metagenomic sequencing of corneal samples on the Illumina NextSeq 500 [<a href="#ref-3">3</a>]. The protocol generated libraries from as little as 1.9 ng per microliter of DNA, demonstrating the sensitivity of shotgun approaches for clinical samples [<a href="#ref-3">3</a>].

Sequencing depth should be determined based on the research question and the expected complexity of the microbial community. Shallow sequencing at approximately 1 Gb per sample provides taxonomic profiles that resemble deep sequencing results for dominant taxa [<a href="#ref-4">4</a>]. Deep sequencing is required for detecting rare taxa, reconstructing genomes, and profiling functional genes comprehensively.

Bioinformatics Analysis Options

The bioinformatics pipeline for shotgun metagenomics includes quality control, taxonomic classification, functional annotation, and optionally assembly. Multiple analysis platforms are available, each with different strengths. Bioconductor provides packages for reproducible genomic analysis, including metagenomic workflows [<a href="#ref-5">5</a>]. The Galaxy Training Network offers accessible workflow training and analysis tutorials for metagenomics [<a href="#ref-6">6</a>]. The nf-core community provides standardized pipelines with documented usage and configuration options [<a href="#ref-12">12</a>].

Researchers should select analysis tools based on their familiarity, the specific research question, and the sequencing platform used. Read-based classification works well for Illumina data and for well-characterized communities. Assembly-based approaches are necessary for strain-level analysis and for discovering novel organisms. Hybrid approaches that combine short and long reads can leverage the strengths of both platforms.

Options and Trade-offs in Platform Selection

Short-Read Only Approaches

Illumina-only shotgun metagenomics remains the most common approach because of its established accuracy, mature analysis ecosystem, and cost-effectiveness for large studies. The approach works well for community profiling, functional gene detection, and antimicrobial resistance surveillance. The primary limitation is assembly fragmentation, which restricts strain-level analysis and the recovery of complete genomes.

For clinical applications, short-read shotgun sequencing can provide rapid antimicrobial resistance prediction. A study using machine learning on Klebsiella pneumoniae isolates achieved high accuracy for resistance prediction and reduced turnaround time compared to culture-based methods [<a href="#ref-7">7</a>]. The approach predicted antimicrobial susceptibility results with a mean turnaround time of approximately 18 hours, compared to approximately 60 hours for traditional methods [<a href="#ref-7">7</a>].

Long-Read Only Approaches

Nanopore-only shotgun metagenomics provides real-time analysis and long reads for contiguous assembly. The approach is valuable for field applications, outbreak investigations, and wastewater surveillance where rapid results are needed [<a href="#ref-9">9</a>]. The higher error rate requires careful analysis strategies, including error correction and validation of variant calls.

PacBio-only shotgun metagenomics provides high-accuracy long reads but at higher cost. The approach is best suited for projects that require complete genome reconstruction and detailed strain analysis. The cost limits the number of samples that can be sequenced, making it less practical for large cohort studies.

Hybrid Approaches

Hybrid assembly combines short and long reads to leverage the strengths of both platforms. Short reads provide high accuracy for error correction, while long reads provide contiguity for assembly. Hybrid approaches can produce complete microbial genomes from complex communities, enabling detailed genomic analysis.

The choice of hybrid strategy depends on the research question and available resources. Some projects may benefit from sequencing a subset of samples with long reads and the full cohort with short reads. Others may require long-read sequencing of all samples to achieve the desired resolution.

Observations and Measurements for Platform Assessment

Metrics for Evaluating Sequencing Performance

Researchers should track several metrics to evaluate sequencing performance and data quality. Read length distribution indicates whether the library preparation preserved DNA fragment sizes. Base quality scores indicate the accuracy of individual base calls. Throughput per run determines how many samples can be sequenced at the desired depth.

For taxonomic profiling, the proportion of reads classified to each taxonomic level indicates the sensitivity of the assay. The number of species detected and the diversity estimates provide measures of community coverage. For functional profiling, the number of genes predicted and the completeness of metabolic pathways indicate the depth of functional information.

Mock Community Controls

Mock communities with known composition provide essential controls for evaluating platform performance. A study using complicated mock microbiomes with more than 60 human gut bacterial species demonstrated that shallow whole-metagenome shotgun sequencing at 1 Gb provided outputs that highly resembled deep sequencing data at both genus and species levels [<a href="#ref-4">4</a>]. The same study found that 16S amplicon results showed poor consistency with whole-metagenome shotgun data [<a href="#ref-4">4</a>].

Mock community controls should be included in every sequencing run to assess batch effects and platform-specific biases. The controls should reflect the expected complexity of the study samples, including both even and varied abundance distributions [<a href="#ref-4">4</a>].

Replicates and Technical Variation

Technical replicates assess the reproducibility of the entire workflow, from DNA extraction through sequencing and analysis. A study of human fecal microbiomes established the reproducibility of shotgun methods with extensive multiplexing [<a href="#ref-1">1</a>]. The study compared multiple sequencing methods and platforms, demonstrating that consistent results could be obtained across technical replicates [<a href="#ref-1">1</a>].

Biological replicates assess the variation between samples from the same population or condition. The number of biological replicates required depends on the expected effect size and the variability of the community. Researchers should perform power calculations based on pilot data to determine the appropriate sample size.

Records and Documentation Requirements

Metadata Collection

Comprehensive metadata collection is essential for interpreting shotgun metagenomic data. Sample metadata should include collection date, location, sample type, storage conditions, and any relevant clinical or environmental information. DNA extraction metadata should include the extraction method, input amount, and DNA yield. Sequencing metadata should include the platform, instrument, chemistry version, and sequencing depth.

The NCBI provides databases and search systems for depositing and accessing sequence data [<a href="#ref-13">13</a>]. Researchers should deposit raw sequencing data and associated metadata in public repositories to enable reproducibility and secondary analysis. The EMBL-EBI provides training and data resources for bioinformatics analysis [<a href="#ref-14">14</a>].

Quality Control Records

Quality control records should document the performance of each sequencing run and the results of quality assessments. These records enable troubleshooting when problems arise and provide evidence of data quality for publications and regulatory submissions.

Key quality metrics include the number of reads generated, the proportion of reads passing quality filters, the average read length, and the base quality scores. For taxonomic analysis, the proportion of reads classified and the distribution of classifications across taxa provide quality indicators. For assembly-based analysis, the number of contigs, the N50 length, and the completeness of recovered genomes indicate assembly quality.

Analysis Documentation

Analysis documentation should record the software versions, parameters, and reference databases used for each analysis step. This documentation enables reproducibility and facilitates troubleshooting when results are unexpected. The nf-core documentation provides standards for pipeline usage and configuration that support reproducible analysis [<a href="#ref-12">12</a>]. The Carpentries lessons provide foundational training in computing, data management, and version control that support reproducible research practices [<a href="#ref-15">15</a>].

Common Failure Patterns and Troubleshooting

Low DNA Yield or Quality

Low DNA yield is a common problem in shotgun metagenomics, particularly for low-biomass samples. The problem manifests as insufficient sequencing depth, poor library complexity, or failed sequencing runs. Solutions include optimizing the extraction method, increasing the input amount, or using whole-genome amplification.

A study of aerosol samples found that shotgun metagenomic sequencing requires relatively large DNA inputs, which is challenging for low-biomass air environments [<a href="#ref-11">11</a>]. The study demonstrated that a custom extraction method improved DNA yields from filter-collected air samples by isolating DNA from the entire filter extract [<a href="#ref-11">11</a>].

Contamination

Contamination can arise from extraction reagents, laboratory environments, or sequencing instruments. A study of extraction methods found that the phenol:chloroform approach showed evidence of contamination in negative controls [<a href="#ref-10">10</a>]. The study recommended including negative controls in every extraction batch to detect contamination.

Contamination is particularly problematic for low-biomass samples, where contaminating DNA can represent a large proportion of the total sequencing output. Researchers should include extraction blanks and sequencing controls to identify and quantify contamination.

Taxonomic Misclassification

Taxonomic misclassification can arise from errors in reference databases, insufficient read length, or high error rates. Short reads may not contain enough unique information to distinguish closely related species. High error rates can cause reads to match the wrong reference genome.

The choice of reference database and classification algorithm affects the accuracy of taxonomic assignments. Researchers should use curated reference databases and validate classifications with complementary methods. For clinical samples, confirmation of pathogen identification may require additional testing.

Assembly Fragmentation

Assembly fragmentation occurs when repetitive regions or closely related strains prevent contiguous reconstruction. The problem is more severe with short-read data and in complex communities. Solutions include using long-read data, increasing sequencing depth, or using binning approaches to separate genomes from different organisms.

A review of soil microbiome studies found that shotgun metagenomics reveals systematic biases in both short and long-read assembly that depend on population diversity within the sample [<a href="#ref-8">8</a>]. The review emphasized that method choice should be framed as an important part of study design, with the biases of the chosen method acknowledged [<a href="#ref-8">8</a>].

Limitations and Interpretation Boundaries

Detection Limits and Sensitivity

Shotgun metagenomics has detection limits that depend on sequencing depth, community complexity, and the abundance of target organisms. Rare taxa may be missed entirely, particularly in complex communities where dominant organisms consume most of the sequencing capacity. The sensitivity for detecting specific pathogens depends on their abundance relative to the total community.

A study of corneal infections demonstrated that shotgun sequencing could detect herpes simplex virus type 1 from a sample with 1.9 ng per microliter of DNA [<a href="#ref-3">3</a>]. The study highlighted the hypothesis-free nature of shotgun sequencing, which identifies the full taxonomic and functional profile of an organism [<a href="#ref-3">3</a>].

Reference Database Dependence

Taxonomic classification depends on the completeness and accuracy of reference databases. Novel organisms with no close relatives in the database may remain unclassified or be misclassified. Functional annotation similarly depends on the availability of characterized genes and pathways.

The NCBI provides comprehensive sequence databases and search systems for taxonomic and functional analysis [<a href="#ref-13">13</a>]. Researchers should use the most current versions of reference databases and document the versions used in their analyses.

Strain-Level Resolution

Strain-level resolution requires sufficient genetic variation to distinguish closely related organisms. Short reads may not provide enough information to resolve strains, particularly in complex communities. Long reads and complete genome assemblies provide the best resolution but require higher sequencing depth and more complex analysis.

A study of Pseudomonas aeruginosa clinical isolates used shotgun whole-genome sequencing to compare genomic divergence, identify antimicrobial resistance genes, and assess phylogenetic relatedness [<a href="#ref-16">16</a>]. The study demonstrated the value of whole-genome sequencing for strain-level characterization of clinically important pathogens [<a href="#ref-16">16</a>].

Safety and Regulatory Context

Clinical Diagnostic Applications

Shotgun metagenomics has potential applications in clinical diagnostics, including pathogen detection and antimicrobial resistance prediction. The approach can identify pathogens without prior knowledge of the suspected organism, making it valuable for difficult-to-diagnose infections [<a href="#ref-3">3</a>]. The approach can also predict antimicrobial susceptibility from sequencing data, potentially reducing the time to appropriate treatment [<a href="#ref-7">7</a>].

Clinical applications require validation of the entire workflow, including sample collection, DNA extraction, sequencing, and analysis. The accuracy and reproducibility of the method must be established before clinical use. Regulatory requirements vary by jurisdiction and application.

Antimicrobial Resistance Surveillance

Shotgun metagenomics provides a powerful tool for antimicrobial resistance surveillance in clinical, environmental, and agricultural settings. The approach detects resistance genes directly from community DNA, without requiring culture of individual organisms. The approach can also attribute resistance genes to specific species using species-specific markers [<a href="#ref-7">7</a>].

A study of Klebsiella pneumoniae developed a genotypic antimicrobial resistance testing method using metagenomic sequencing data [<a href="#ref-7">7</a>]. The method used machine learning to identify genetic resistance determinants and achieved high accuracy for resistance prediction [<a href="#ref-7">7</a>].

Wastewater Surveillance

Wastewater surveillance using shotgun metagenomics can detect pathogens and antimicrobial resistance genes in community populations. The approach provides early warning of disease outbreaks and supports public health decision-making. A study of wastewater samples in Central India demonstrated the utility of shotgun metagenomics for virus characterization at the strain level [<a href="#ref-9">9</a>].

The study used a novel nanoweb membrane-based sample enrichment method followed by shotgun whole-genome metagenomics on Nanopore and Ion Torrent platforms [<a href="#ref-9">9</a>]. The method was field-deployable and cost-effective, supporting its use in routine surveillance programs [<a href="#ref-9">9</a>].

Professional Escalation Criteria

When to Seek Specialized Support

Researchers should seek specialized support when encountering problems that exceed their local expertise. These situations include persistent contamination issues, unexpected taxonomic results, assembly failures, or regulatory questions. Bioinformatics support may be needed for complex analyses or for implementing reproducible workflows.

The Galaxy Training Network provides accessible workflow training and analysis tutorials that can help researchers build their analysis skills [<a href="#ref-6">6</a>]. The nf-core documentation provides standards for community pipelines that support reproducible analysis [<a href="#ref-12">12</a>]. The Carpentries lessons provide foundational training in computing and data management [<a href="#ref-15">15</a>].

When to Consider Alternative Approaches

Researchers should consider alternative approaches when shotgun metagenomics does not answer the research question. Targeted approaches such as hybridization capture may be more appropriate for detecting specific organisms or genes at lower cost [<a href="#ref-17">17</a>]. Amplicon sequencing may be sufficient for community profiling when species-level resolution is not required.

A study comparing shotgun sequencing and rpoB metabarcoding for taxonomic profiling of bacterial communities found that the two approaches have different strengths [<a href="#ref-18">18</a>]. The choice between approaches should be based on the research question, the expected community composition, and the available resources.

When to Escalate to Clinical or Regulatory Authorities

Clinical or regulatory escalation is required when shotgun metagenomic results have implications for patient care, public health, or regulatory compliance. Suspected outbreaks, notifiable diseases, or unexpected pathogens should be reported to appropriate authorities. Antimicrobial resistance findings with clinical implications should be communicated to treating clinicians.

The decision to escalate should be documented, including the rationale and the actions taken. Communication with relevant stakeholders should be timely and clear, with appropriate attention to data confidentiality and regulatory requirements.

A Practical Decision Framework for Platform Selection Based on Project Constraints

Selecting a sequencing platform requires more than comparing technical specifications. The decision must account for the specific constraints of each project, including sample throughput, budget per sample, required turnaround time, and the complexity of the microbial community under study. This section provides a structured framework for matching platform characteristics to project requirements, with emphasis on the trade-offs that matter most in practice.

Step 1: Define the Primary Biological Question

The first step in platform selection is to define the primary biological question that the sequencing data must answer. This question determines the minimum acceptable read length, accuracy, and depth. A study focused on species-level taxonomic profiling of well-characterized communities may succeed with short-read data alone. A study aimed at recovering complete microbial genomes from complex environmental samples requires long-read data for contiguous assembly.

The distinction between read-based and assembly-based analysis is central to this decision. Read-based approaches classify individual reads against reference databases and work well for communities where reference genomes exist for most members. Assembly-based approaches reconstruct longer genomic fragments and are necessary for discovering novel organisms, resolving strain-level variation, and characterizing mobile genetic elements. A review of whole-metagenome shotgun sequencing studies identifies assembly, community profiling, and functional profiling as three major research areas, each with different data requirements [<a href="#ref-2">2</a>].

Step 2: Assess Sample Characteristics and DNA Input

Sample type and expected DNA yield constrain platform choice more than any other factor. Low-biomass samples such as aerosols, sputum, and vacuumed dust present challenges for all platforms but are particularly difficult for long-read approaches that require high molecular weight DNA. A systematic evaluation of extraction methods found that the proportion of human reads varied dramatically by sample type, with stool having the lowest proportion of human reads at 0.1 percent, compared to dust at 44.1 percent and sputum at 80 percent [<a href="#ref-10">10</a>]. This variation affects the effective microbial sequencing depth and must be considered when calculating required throughput.

For low-biomass samples, the extraction method itself introduces variability that can affect downstream results. The same study found that DNA yield varied by extraction method, with phenol:chloroform and Promega approaches producing the highest yields, and that the extraction method affected microbial community structure and functional profiles [<a href="#ref-10">10</a>]. A custom multi-component DNA isolation method optimized for aerosol samples improved DNA yields from filter-collected air samples by isolating DNA from the entire filter extract [<a href="#ref-11">11</a>]. Researchers working with low-biomass samples should validate their extraction method using mock community controls before committing to a sequencing platform.

Step 3: Calculate Required Sequencing Depth

Sequencing depth must be calculated based on the expected community complexity and the detection limits required by the research question. Shallow whole-metagenome shotgun sequencing at approximately 1 Gb per sample provides taxonomic assignments at genus and species levels that closely resemble deep sequencing results, while remaining more cost-effective for large cohort studies [<a href="#ref-4">4</a>]. This finding is particularly relevant for longitudinal studies with hundreds or thousands of samples, where the cost of deep sequencing would be prohibitive.

For functional profiling and rare taxon detection, deeper sequencing is required. The relationship between depth and detection sensitivity is not linear, and the optimal depth depends on the abundance distribution of the community. Communities with many rare taxa require deeper sequencing to capture their diversity. Communities dominated by a few abundant organisms may be adequately characterized at lower depths.

Step 4: Evaluate Turnaround Time Requirements

Turnaround time requirements vary by application and can be the deciding factor in platform selection. Clinical applications that require rapid results for treatment decisions benefit from platforms with short run times and real-time analysis capabilities. A study of antimicrobial resistance prediction in Klebsiella pneumoniae demonstrated that shotgun metagenomic sequencing could provide genotypic antimicrobial susceptibility testing results with a mean turnaround time of approximately 18 hours, compared to approximately 60 hours for traditional culture-based methods [<a href="#ref-7">7</a>].

Nanopore sequencing provides real-time data generation, allowing researchers to begin analysis before the run completes. This capability is valuable for outbreak investigations and wastewater surveillance where early detection is critical. A study of wastewater samples demonstrated the utility of Nanopore-based shotgun metagenomics for strain-level virus characterization, including Influenza A and SARS-CoV-2 [<a href="#ref-9">9</a>]. PacBio sequencing requires the full run to complete before analysis can begin, which may be a limitation for time-sensitive applications.

Step 5: Compare Cost Per Sample Across Platforms

Cost per sample is a primary constraint for most research projects, and the optimal platform depends on the number of samples and the required depth per sample. Illumina platforms offer the lowest cost per gigabase, making them the most economical choice for large cohort studies with shallow sequencing depth. The cost advantage of short-read platforms is well established, and shallow whole-metagenome shotgun sequencing represents a cost-efficient alternative to 16S amplicon sequencing for large-scale microbiome studies [<a href="#ref-4">4</a>].

Long-read platforms have higher cost per gigabase but may reduce overall project cost by eliminating the need for additional validation experiments. Complete genome assemblies from long-read data can provide strain-level resolution that would require extensive additional analysis with short-read data alone. The cost comparison should include also sequencing costs but also the costs of library preparation, bioinformatics analysis, and any follow-up validation experiments.

Step 6: Consider Bioinformatics Capacity and Expertise

The bioinformatics requirements of each platform differ substantially, and the available expertise within the research team should inform platform selection. Illumina short-read data are compatible with the widest range of analysis tools and workflows. Bioconductor provides packages for reproducible genomic analysis, including metagenomic workflows [<a href="#ref-5">5</a>]. The Galaxy Training Network offers accessible workflow training and analysis tutorials for metagenomics [<a href="#ref-6">6</a>]. The nf-core community provides standardized pipelines with documented usage and configuration options [<a href="#ref-12">12</a>].

Long-read data require specialized analysis tools for error correction, assembly, and variant calling. The higher error rates of Nanopore data require careful quality filtering and validation of results. Researchers without prior experience in long-read analysis should factor in the time required for training and pipeline development. The Carpentries lessons provide foundational training in computing, data management, and version control that support reproducible research practices [<a href="#ref-15">15</a>].

Step 7: Document the Decision and Its Rationale

The platform selection decision should be documented with the rationale for each choice. This documentation supports reproducibility and provides a reference for troubleshooting if problems arise during the project. The documentation should include the primary biological question, sample characteristics, required depth, turnaround time requirements, cost constraints, and bioinformatics capacity.

The documentation should also record the expected limitations of the chosen platform and the strategies for mitigating those limitations. For example, a project using short-read data for assembly should document the expected fragmentation and the plan for validating assemblies with complementary methods. A project using long-read data should document the expected error rates and the plan for error correction and validation.

A Structured Comparison Matrix for Project Planning

The following comparison matrix provides a structured approach to evaluating platform options against project-specific requirements. Each project should be scored across the dimensions of read length, accuracy, throughput, cost, turnaround time, and bioinformatics complexity. The scoring should reflect the specific requirements of the project instead of general platform characteristics.

The matrix should be used as a starting point for project planning, not as a definitive prescription. The specific requirements of each project will determine which dimensions are most important and how the trade-offs should be weighted.

Records and Measurements for Platform Validation

Pre-Project Validation Records

Before committing to a sequencing platform for a full project, researchers should conduct a validation study using representative samples and mock community controls. The validation study should assess DNA yield, library preparation success, sequencing output, and data quality. The results should be documented and compared against the project requirements.

A study using complicated mock microbiomes with more than 60 human gut bacterial species demonstrated the value of defined communities for evaluating sequencing methods [<a href="#ref-4">4</a>]. The study found that shallow whole-metagenome shotgun sequencing at 1 Gb provided outputs that highly resembled deep sequencing data at both genus and species levels [<a href="#ref-4">4</a>]. The same study found that 16S amplicon results showed poor consistency with whole-metagenome shotgun data, highlighting the importance of method validation [<a href="#ref-4">4</a>].

Run-Level Quality Metrics

Each sequencing run should be documented with standard quality metrics that enable comparison across runs and platforms. These metrics include the number of reads generated, the proportion of reads passing quality filters, the average read length, and the base quality scores. For paired-end short-read data, the insert size distribution should also be recorded.

For taxonomic analysis, the proportion of reads classified to each taxonomic level provides a quality indicator. A low proportion of classified reads may indicate reference database limitations, sequencing errors, or the presence of novel organisms. For assembly-based analysis, the number of contigs, the N50 length, and the completeness of recovered genomes indicate assembly quality.

Longitudinal Tracking of Platform Performance

Research groups that use sequencing platforms across multiple projects should maintain longitudinal records of platform performance. These records enable early detection of instrument drift, reagent lot variations, and protocol changes. The records should include the instrument, chemistry version, library preparation protocol, and quality metrics for each run.

The NCBI provides databases and search systems for depositing and accessing sequence data [<a href="#ref-13">13</a>]. Researchers should deposit raw sequencing data and associated metadata in public repositories to enable reproducibility and secondary analysis. The EMBL-EBI provides training and data resources for bioinformatics analysis [<a href="#ref-14">14</a>].

Common Failure Patterns in Platform Selection

Overestimating the Value of Read Length

Researchers sometimes select long-read platforms based on the assumption that longer reads always produce better results. This assumption is not always correct. For well-characterized communities where reference genomes exist for most members, short-read data can provide accurate species-level taxonomic profiles at lower cost. The marginal benefit of long reads depends on the specific research question and the complexity of the community.

A review of soil microbiome studies found that while long-read and short-read 16S approaches generally converge on dominant taxa and between-sample differences, they disagree substantially on alpha diversity estimates, rare taxon detection, and the relative abundances of entire phyla [<a href="#ref-8">8</a>]. The review emphasized that method choice should be framed as an important part of study design, with the biases of the chosen method acknowledged [<a href="#ref-8">8</a>].

Underestimating the Impact of DNA Extraction

DNA extraction method introduces variability that can affect downstream results, and this variability is sometimes underestimated in platform selection. A systematic evaluation of four extraction methods found that the extraction method affected microbial community structure and functional profiles [<a href="#ref-10">10</a>]. The same study found that only the phenol:chloroform approach showed evidence of contamination in negative controls [<a href="#ref-10">10</a>].

Researchers should validate their extraction method using mock community controls before committing to a sequencing platform. The extraction method should be documented and kept consistent across all samples in a study to minimize technical variation.

Ignoring Bioinformatics Bottlenecks

The bioinformatics requirements of each platform differ substantially, and researchers sometimes underestimate the time and expertise required for data analysis. Long-read data require specialized analysis tools and careful validation of results. The higher error rates of Nanopore data require quality filtering and error correction strategies.

Researchers should assess their bioinformatics capacity before selecting a platform and factor in the time required for training and pipeline development. The Galaxy Training Network provides accessible workflow training and analysis tutorials for metagenomics [<a href="#ref-6">6</a>]. The nf-core documentation provides standards for community pipelines that support reproducible analysis [<a href="#ref-12">12</a>].

Failing to Plan for Validation Experiments

Platform selection should include a plan for validating the results, particularly for clinically important findings or novel discoveries. Validation may involve complementary sequencing methods, PCR assays, or culture-based confirmation. The validation plan should be documented and budgeted for in the project plan.

A study of corneal infections demonstrated that shotgun sequencing could confirm a diagnosis of herpes simplex virus type 1 that was suspected by conventional methods [<a href="#ref-3">3</a>]. The study highlighted the hypothesis-free nature of shotgun sequencing, which identifies the full taxonomic and functional profile of an organism [<a href="#ref-3">3</a>]. The validation of clinically important findings is essential before treatment decisions are made.

Escalation Criteria for Platform-Related Problems

When to Seek Bioinformatics Support

Researchers should seek specialized bioinformatics support when encountering problems that exceed their local expertise. These situations include persistent assembly failures, unexpected taxonomic results, or difficulties implementing reproducible workflows. The nf-core documentation provides standards for community pipelines that support reproducible analysis [<a href="#ref-12">12</a>]. The Carpentries lessons provide foundational training in computing and data management [<a href="#ref-15">15</a>].

When to Reconsider Platform Choice

Researchers should reconsider their platform choice when the data quality or quantity does not meet the project requirements. This situation may arise when the sequencing depth is insufficient for the research question, when the error rate prevents confident taxonomic classification, or when the cost per sample exceeds the project budget.

The decision to switch platforms should be based on documented evidence of the limitations of the current approach. The switch should be planned carefully to maintain consistency across the project, and the reasons for the switch should be documented in the project records.

When to Escalate to Clinical or Regulatory Authorities

Clinical or regulatory escalation is required when shotgun metagenomic results have implications for patient care, public health, or regulatory compliance. Suspected outbreaks, notifiable diseases, or unexpected pathogens should be reported to appropriate authorities. Antimicrobial resistance findings with clinical implications should be communicated to treating clinicians.

The decision to escalate should be documented, including the rationale and the actions taken. Communication with relevant stakeholders should be timely and clear, with appropriate attention to data confidentiality and regulatory requirements.

Frequently Asked Questions

What is the difference between shotgun metagenomics and 16S amplicon sequencing?

Shotgun metagenomics sequences all DNA in a sample, providing information about bacteria, archaea, fungi, viruses, and functional genes. The approach detects bacterial species, increases diversity detection, and improves gene prediction compared to 16S amplicon sequencing [<a href="#ref-1">1</a>]. The 16S method targets a single conserved gene and provides limited taxonomic resolution, typically at the genus level for many organisms.

How much sequencing depth do I need for shotgun metagenomics?

The required depth depends on the research question and the complexity of the microbial community. Shallow sequencing at approximately 1 Gb per sample provides taxonomic assignments at genus and species levels that closely resemble deep sequencing results [<a href="#ref-4">4</a>]. Deep sequencing is required for detecting rare taxa, reconstructing genomes, and comprehensive functional profiling.

Can I combine short-read and long-read sequencing in one project?

Hybrid approaches that combine short and long reads can leverage the strengths of both platforms. Short reads provide high accuracy for error correction, while long reads provide contiguity for assembly. The choice of hybrid strategy depends on the research question and available resources.

How do I choose between Nanopore and PacBio for long-read sequencing?

Nanopore provides real-time analysis, portability, and lower instrument cost, but has a higher per-base error rate. PacBio provides higher accuracy with HiFi reads but costs more per gigabase. The choice depends on whether real-time analysis or accuracy is more important for the research question.

What controls should I include in a shotgun metagenomic study?

Include extraction blanks to detect contamination, mock community controls to assess bias, and technical replicates to assess reproducibility. A study using complicated mock microbiomes demonstrated the value of defined communities for evaluating sequencing methods [<a href="#ref-4">4</a>]. Negative controls should be included in every extraction batch [<a href="#ref-10">10</a>].

How do I validate taxonomic assignments from shotgun metagenomic data?

Validate taxonomic assignments using complementary methods, such as PCR or culture, particularly for clinically important findings. Use curated reference databases and document the versions used. For novel or unexpected findings, consider additional sequencing or targeted assays to confirm the result.

What are the main sources of error in shotgun metagenomic analysis?

Main error sources include DNA extraction bias, sequencing errors, reference database incompleteness, and analysis parameter choices. DNA extraction method affects microbial community structure and functional profiles [<a href="#ref-10">10</a>]. Sequencing errors vary by platform, with short-read platforms having lower error rates than long-read platforms [<a href="#ref-8">8</a>].

How should I report shotgun metagenomic results for publication?

Report the sequencing platform, chemistry version, sequencing depth, and quality metrics. Document the analysis pipeline, including software versions, parameters, and reference databases. Deposit raw data and metadata in public repositories such as NCBI [<a href="#ref-13">13</a>]. Provide sufficient detail to enable reproduction of the analysis by other researchers.

Related Bioinformatics Guides

• How to Choose a Long-Read Sequencing Platform: PacBio vs Oxford Nanopore
• Metagenomics vs Metabarcoding: Choosing the Right Approach for Your Study
• Long-Read Sequencing Cost and Market: What to Expect
• Metagenomics vs Metatranscriptomics: Choosing the Right Approach for Functional Profiling
• Spatial Transcriptomics Platforms Compared: Choosing the Right Technology for Your Study

Related Clinical & Scientific Guides

• A Practical Guide to Detecting Antimicrobial Resistance Genes in Shotgun Metagenomic Data
• Computational Immunology: Modeling the Immune System
• How to Set Hard Filters for Germline Variant Calling: A Practical Guide to GATK Best Practices

References and Further Reading

This article is educational and does not replace validated analysis plans, institutional policy, clinical interpretation, or specialist review.