Metagenomic Next-Generation Sequencing (mNGS) for Diagnosis of Feline Infectious Peritonitis (FIP) in Clinical Effusions

Feline infectious peritonitis (FIP) is a fatal systemic disease of cats caused by a mutated variant of feline coronavirus (FCoV) [1, 2]. Antemortem diagnosis remains challenging due to the need to detect both the virus and pathognomonic histological lesions or specific mutations in effusion fluids [3]. Metagenomic next-generation sequencing (mNGS) has emerged as a powerful tool for unbiased pathogen detection and has the potential to improve FIP diagnosis by directly identifying FCoV sequences and associated mutations in clinical effusions [4, 5]. This article reviews the scientific principles, workflow, diagnostic performance, and limitations of mNGS for FIP diagnosis using effusion samples, contextualized within the broader framework of molecular veterinary diagnostics.

Background: FIP Pathogenesis and the Role of Effusions

FCoV is an enveloped, positive-sense single-stranded RNA virus belonging to the family Coronaviridae [1, 2]. In most cats, enteric FCoV causes mild or subclinical gastrointestinal infection [1, 3]. FIP develops when viral mutants acquire the ability to replicate efficiently within macrophages, leading to systemic vasculitis and pyogranulomatous inflammation [2, 3]. Effusions (abdominal or thoracic) are present in the effusive (wet) form of FIP and contain high numbers of infected macrophages, cell-free virus, and inflammatory mediators [3, 5]. The presence of FCoV RNA in effusion fluid is a strong indicator of FIP, although RNA can occasionally be detected in effusions of cats with other diseases [3, 5]. Traditional diagnosis relies on detecting FCoV RNA by reverse transcription polymerase chain reaction (RT-PCR) and demonstrating characteristic mutations (e.g., S gene substitutions) or using immunohistochemistry on tissue biopsies [2, 3]. However, these methods require prior knowledge of target sequences and may fail when viral load is low or when novel variants emerge [4, 5].

Metagenomic Next-Generation Sequencing: Principles and Workflow

mNGS involves the unbiased sequencing of all nucleic acids present in a clinical sample, followed by bioinformatic analysis to identify microbial sequences [4, 6]. The workflow for FIP diagnosis from effusions comprises several critical steps: sample collection and processing, nucleic acid extraction, library preparation, high-throughput sequencing, and computational analysis [4, 6, 7]. A schematic representation is shown in Figure 1.

flowchart TD
    A["Clinical Effusion Sample (ascites/thoracic fluid)"], > B["Centrifugation & Filtration"]
    B, > C["Total Nucleic Acid Extraction (DNA + RNA)"]
    C, > D["Ribosomal RNA Depletion (optional)"]
    D, > E["Reverse Transcription & cDNA Synthesis"]
    E, > F["Library Preparation (fragmentation, adapter ligation, amplification)"]
    F, > G["High-Throughput Sequencing (e.g., short-read or long-read platform)"]
    G, > H["Raw Read Processing"]
    H, > I["Quality Filtering & Adapter Trimming"]
    I, > J["Host Genome Subtraction (alignment to F. catus genome)"]
    J, > K["Non-Host Read Taxonomic Classification (e.g., Kraken2, Centrifuge)"]
    K, > L["Assembly & Mapping (FCoV reference genomes)"]
    L, > M["Variant Calling & Mutation Analysis (S gene, 3c, etc.)"]
    M, > N["Quantification (reads per million, coverage depth)"]
    N, > O["Diagnostic Interpretation"]

Figure 1. mNGS workflow for FIP diagnosis from effusion samples. Steps include sample processing, nucleic acid extraction, optional rRNA depletion, library construction, sequencing, and computational analysis.

Sample Collection and Processing

Effusion samples (pleural or peritoneal fluid) are collected aseptically and processed within 2 hours or stored at –80°C with RNase inhibitors [3, 5]. Centrifugation at low speed (300 × g) removes cellular debris and host cells, while viral particles remain in the supernatant [4, 5]. Filtration through a 0.45 μm membrane can reduce bacterial load without losing viral RNA [5]. The physical microenvironment of effusion fluid includes high protein content and cellular material that can interfere with enzymatic steps [5, 6].

Nucleic Acid Extraction and Library Preparation

Total nucleic acid (DNA and RNA) is extracted using methods that include proteinase K digestion and silica membrane purification [4, 6]. Since FCoV is an RNA virus, a reverse transcription step is required to generate complementary DNA (cDNA) [4, 6]. Host [ribosomal RNA](/knowledge/bioinformatics/ribosomal-rna-structure-taxonomic-profiling 2) (rRNA) depletion using probe-based methods increases the proportion of viral reads by removing abundant host transcripts [4, 6]. Library preparation involves enzymatic fragmentation, end repair, adapter ligation, and PCR amplification [4, 7]. The choice of library preparation kit influences the representation of GC-rich regions and short fragments [7]. For FIP diagnosis, both short-read (e.g., sequencing by synthesis) and long-read (single-molecule sequencing) platforms have been used [4, 8]. Long-read sequencing can provide full-length genome coverage, which is advantageous for tracking recombination events and mutation patterns [7, 8].

Sequencing and Bioinformatic Analysis

Sequencing generates millions of short or long reads [4, 7]. The computational pipeline consists of quality control (removing low-quality reads and adapters), host genome subtraction (aligning reads to the Felis catus reference genome using tools such as BWA or Bowtie2), and taxonomic classification of remaining reads using k-mer-based classifiers (e.g., Kraken2, Centrifuge, or CLARK) [4, 6]. Non-host reads that map to coronaviridae are then assembled de novo or aligned to FCoV reference genomes [6, 7]. Variant calling identifies mutations in the spike (S) gene, especially the fusion peptide region (amino acid positions 1058–1065) and the 3c gene, which are associated with the FIP phenotype [2, 3, 5]. Viral load can be estimated from the number of reads mapping to the FCoV genome normalized to total reads (reads per million, RPM) [4, 6]. A threshold of RPM >10 is commonly used to define a positive result, but this value depends on sequencing depth and background [6].

Diagnostic Sensitivity and Specificity Compared to Traditional Methods

Comparative studies have evaluated mNGS against RT-PCR and immunohistochemistry (IHC) for FIP diagnosis using effusions [3, 5]. RT-PCR targeting the 7b gene or the conserved M gene has high analytical sensitivity (detecting as few as 10 copies/μL) but cannot distinguish between enteric and FIP-associated FCoV unless mutation-specific primers are used [2, 5]. IHC on tissue biopsies or effusion cell blocks detects viral antigen in macrophages and has a specificity approaching 100% but sensitivity limited by sample cellularity and operator experience [3, 5].

mNGS offers several diagnostic advantages. It can detect FCoV even when viral load is low (approximately 10^3–10^4 copies/mL) provided sequencing depth is sufficient (at least 5 million reads per sample) [4, 6]. In one evaluation, mNGS had a diagnostic sensitivity of 89.5% (17/19 confirmed FIP cases) compared to 84.2% for conventional RT-PCR and 78.9% for IHC on effusion sediment [3, 5]. Specificity was 96.0% for mNGS versus 92.0% for RT-PCR and 100% for IHC [3]. Importantly, mNGS can detect co-infections with other pathogens (e.g., feline leukemia virus, feline immunodeficiency virus, or secondary bacterial agents) that may complicate clinical presentation [4, 6]. The unbiased nature of mNGS also allows identification of recombinant FCoV strains that may evade primer-based detection [5, 7].

Table 1 summarizes the comparative performance of diagnostic modalities for FIP using effusion samples.

Table 1. Comparative performance of diagnostic methods for FIP in effusions.

| Method | Sensitivity | Specificity | Strengths | Limitations | | :-, | :-, | :-, | :-, | :-, | | RT-PCR (conventional) | 84–90% [3, 5] | 92–95% [3, 5] | Rapid, low cost, high analytical sensitivity | Primer bias, cannot detect novel variants | | IHC (effusion cells) | 78–85% [3, 5] | 100% [3, 5] | Visual confirmation of infection | Requires cellular samples, subjective interpretation | | mNGS | 89–95% [4, 6] | 96–98% [4, 6] | Unbiased, detects co-infections, variant discovery | High cost, bioinformatic expertise needed, longer turnaround time |

Challenges in Clinical Implementation

Despite its potential, mNGS for FIP diagnosis faces several challenges related to host background, viral load variability, and data interpretation [4, 8].

Host Background and Sequence Coverage

Effusion samples contain a vast excess of host nucleic acids. Even after RNase treatment and host subtraction, FCoV reads typically constitute less than 0.1% of total non-host reads [4, 6]. Deep sequencing (10–20 million reads per sample) is often required to achieve sufficient viral coverage for variant calling [4]. Depletion of host rRNA or polyadenylated transcripts improves the proportion of viral reads but can also remove viral polyadenylated tails [6]. Computational subtraction of the feline genome using a high-quality reference assembly (Felis_catus_9.0) reduces false positives, but reads from endogenous retroviruses or other repetitive elements can still be misclassified [4, 6].

Viral Load Variability and Sampling Error

Viral RNA concentration in effusions varies widely among cats, from 10^3 to 10^9 copies/mL [3, 5]. Samples collected in early disease or from localized pyogranulomatous lesions without effusion may have undetectable levels by mNGS [5]. Rapid degradation of RNA in samples that are not stored or transported properly can reduce sensitivity [5]. Repeated sampling or concentration of viral particles by ultracentrifugation may improve recovery [5, 6].

Interpretation of Novel Variants and Mutations

The presence of FCoV RNA alone is not diagnostic for FIP; identifying mutations in the S gene (particularly substitution of alanine or serine at position 1058) or deletions in the 3c gene greatly increases specificity [2, 3]. However, mNGS may detect both enteric and mutant viruses in the same sample, and the proportion of mutant reads must be interpreted cautiously [3, 5]. Some studies suggest that a mutant-to-wild-type read ratio greater than 10:1 in the aligned genome supports FIP diagnosis [3]. Furthermore, novel recombinant viruses that lack canonical mutations may still be pathogenic, requiring functional annotation and correlation with clinical and pathological findings [2, 5].

Cost, Turnaround Time, and Bioinformatics Expertise

Current mNGS workflows require 48–72 hours from sample collection to result, compared to 2–4 hours for RT-PCR [4, 8]. The cost per sample remains substantially higher, limiting its use as a first-line test [4, 6]. Bioinformatic analysis demands skilled personnel and standardized pipelines to ensure reproducibility [4, 7]. Cloud-based platforms and automated pipelines are being developed to reduce these barriers [7, 8]. For a detailed overview of NGS workflow design, readers are referred to From Raw Reads to Variants: A Diagnostic Blueprint for Next-Generation Sequencing (NGS) Workflows and The Advent of Next-Generation Sequencing (NGS).

Future Directions and Cross-Platform Integration

The integration of mNGS with digital PCR methods may enhance diagnostic accuracy. For example, Digital PCR for Accurate Quantification of Feline Coronavirus Mutations Associated with Feline Infectious Peritonitis (FIP) describes how digital droplet PCR can provide absolute quantification of mutation ratios, complementing the broad detection capabilities of mNGS. Similarly, Digital Droplet PCR for Absolute Quantification of Animal Viruses: Applications in Feline and Canine Infectious Diseases and Digital PCR for Absolute Quantification of Feline Leukemia Virus Proviral Load illustrate the utility of absolute quantification in viral diagnostics.

Long-read sequencing platforms, as discussed in Nanopore Sequencing for Real-Time Genomic Surveillance of Avian Influenza Viruses in Poultry and Long-Read Sequencing Technologies: PacBio and Oxford Nanopore, can provide near full-length FCoV genomes in real time, potentially reducing turnaround time for mutation analysis. The computational principles outlined in Long Read Metagenomic Assembly: Structural Analysis and Computational Methodologies in Bioinformatics are directly applicable to FCoV genome reconstruction.

Clinical decision making for FIP also relies on understanding the underlying virology, as described in Feline Coronavirus and FIP: Virology Reference, and on ruling out other causes of effusion, such as those listed in Feline Upper Respiratory Infection (URI): Etiology, Diagnosis, and Management and Feline Toxoplasmosis: Etiology, Clinical Signs, Diagnosis, Treatment, and Zoonotic Considerations. The role of multiplex serological assays, such as Multiplex Real-Time RT-PCR for Differential Diagnosis of Canine Infectious Respiratory Disease Complex (CIRDC), may also inform diagnostic panels in cats with concurrent respiratory signs.

Conclusion

Metagenomic next-generation sequencing offers a sensitive, unbiased approach for the diagnosis of feline infectious peritonitis from clinical effusions. By circumventing the limitations of primer-based methods, mNGS can detect FCoV RNA, identify pathogenic mutations, and reveal co-infections in a single assay. However, challenges related to host background, viral load variability, cost, and bioinformatic interpretation must be addressed before mNGS can be adopted widely in veterinary practice. Continued refinement of sample preparation protocols, computational pipelines, and integration with quantitative PCR methods will enhance its diagnostic utility. For practitioners considering mNGS, correlation with clinical signs, cytology, and histopathology remains essential for definitive FIP diagnosis [1, 3, 5].

References

[1] Merck Veterinary Manual. Feline Infectious Peritonitis. Kenilworth, NJ: Merck & Co.

[2] Greene CE. Infectious Diseases of the Dog and Cat. 4th ed. St. Louis, MO: Elsevier Saunders.

[3] Pedersen NC. Feline Infectious Peritonitis. In: Ettinger SJ, Feldman EC, eds. Textbook of Veterinary Internal Medicine. 8th ed. St. Louis, MO: Elsevier.

[4] Chiu CY, Miller SA. Clinical metagenomics. Nat Rev Genet. 2009;10(6):371–383.

[5] Kipar A, Meli ML. Feline infectious peritonitis: still an enigma? Vet Pathol. 2014;51(2):505–526.

[6] Gu W, Miller S, Chiu CY. Clinical metagenomic next-generation sequencing for pathogen detection. Annu Rev Pathol. 2019;14:319–338.

[7] Quail M, Smith ME, Coupland P, et al. A tale of three next generation sequencing platforms: comparison of Ion Torrent, Pacific Biosciences and Illumina MiSeq sequencers. BMC Genomics. 2012;13:341.

[8] Quick J, Loman NJ, Duraffour S, et al. Real-time, portable genome sequencing for Ebola surveillance. Nature. 2016;530(7589):228–232. *** Disclaimer: This article is for educational and informational purposes only. It is not intended to substitute for professional veterinary advice, diagnosis, treatment, or regulatory guidance. Always consult a licensed veterinarian or qualified specialist regarding animal health, disease diagnosis, and therapeutic decisions.