Co-infections and cryptic pathogens uncovered by metatranscriptomics in New Zealands severe acute respiratory infections

By applying metatranscriptomic sequencing to 300 PCR-negative severe acute respiratory infection samples in New Zealand, this study uncovered a diverse array of co-infecting viral, bacterial, and fungal pathogens missed by routine diagnostics, demonstrating the critical value of genomic approaches for improving diagnostic accuracy and public health surveillance.

The Gist

Imagine your lungs are a bustling city. Usually, when someone gets a severe respiratory infection (like a bad flu or pneumonia), doctors act like traffic cops. They have a specific list of "wanted" criminals (viruses like Flu, RSV, or SARS-CoV-2) and use a high-tech scanner (PCR tests) to check if any of those specific criminals are in the city.

If the scanner says "No criminals found," the patient is often left with a diagnosis of "mystery illness." The doctors know the city is under attack, but they don't know who the attackers are.

This paper is about a team of researchers in New Zealand who decided to stop using the "wanted list" scanner and instead used a super-powered, all-seeing drone (called metatranscriptomics) to fly over the city and take a video of everything that was actually active and moving.

Here is the breakdown of what they found, using simple analogies:

1. The "Mystery" Patients

The researchers looked at 300 patients who were in the hospital with severe breathing problems. The standard hospital tests had already told them, "We don't know what's wrong; no common viruses detected."

The researchers took a sample from the back of the patients' noses and used their "drone" to read the genetic code of every single living thing that was actively working (transcribing RNA) in that sample.

2. The Big Reveal: 43% Were Solved!

The drone found that in 43% of these "mystery" cases, there was indeed a culprit. The standard tests had missed them because they weren't on the "wanted list."

• The Sneaky Viruses: They found common cold viruses (Rhinoviruses) and seasonal coronaviruses that the hospital tests simply didn't look for. It's like the traffic cop only checking for red cars, while the city was being attacked by blue and green cars.
• The "Hidden" Villains: They also found viruses that are rarely tested for, like Measles, Parainfluenza, and even some herpes viruses.
• The Bacteria and Fungi: It wasn't just viruses. They found bacteria (like Strep pneumoniae) and fungi that were actively causing trouble, which standard viral tests would never catch.

3. The "Team-Up" (Co-infections)

One of the most interesting findings was that the attackers often worked in teams.

• The Analogy: Imagine a bank robbery. Usually, you think it's just one guy with a mask. But the researchers found that often, it was a whole gang: a virus breaking down the doors, while bacteria and fungi rushed in to steal the loot.
• The Stats: In 26% of the cases where they found a pathogen, there were actually multiple different types of germs attacking at the same time. The standard tests usually only catch the loudest one (or none at all), missing the whole gang.

4. Why Did the Standard Tests Fail?

The researchers checked the "wanted list" (the PCR primers) against the viruses they found. They realized the tests should have worked, but they didn't.

• The Metaphor: It's like a metal detector that is tuned to find gold coins. If someone is stealing silver coins, the detector stays silent. Or, the detector is looking for a specific shape of coin, but the thief has a slightly different shape. The technology wasn't broken; it was just too narrow-minded.

5. The "Criminal Record" (Virulence and Resistance)

The "drone" didn't just identify who was there; it also read their "criminal record."

• Virulence Factors: They found genes that act like weapons (toxins) or tools to break into cells. This told them these germs weren't just hanging out; they were actively attacking.
• Antibiotic Resistance: They found "shield genes" that tell bacteria how to ignore antibiotics. This is crucial because if a doctor prescribes the wrong antibiotic, these bacteria will just laugh and keep attacking.

6. Who Gets Hit Hardest?

The study looked at the demographics:

• Age: The "microbial city" looked different depending on the patient's age. Children had different "gangs" of germs compared to the elderly.
• Time: The types of germs changed from year to year, showing that the "criminal landscape" shifts over time.

The Bottom Line

This paper is a wake-up call for the medical world.

The Old Way: "We have a list of 10 bad guys. If we don't see them, we say 'we don't know what's wrong.'"
The New Way (Metatranscriptomics): "Let's look at the whole city. We found 10 bad guys you didn't know about, and they are often working in gangs."

Why it matters:
If doctors start using this "all-seeing drone" approach, they can:

1. Stop guessing why patients are sick.
2. Prescribe the exact right medicine (avoiding useless antibiotics).
3. Catch outbreaks of rare or new viruses before they spread.

In short, the researchers opened the blinds in a dark room and realized the monsters were there all along; we just needed a better flashlight to see them.

Technical Summary

1. Problem Statement

Severe Acute Respiratory Infections (SARI) are a leading cause of global hospitalization and mortality. A significant clinical challenge is the high rate of "undiagnosed" SARI cases where standard diagnostic panels fail to identify an etiological agent.

• Limitations of Current Diagnostics: Routine clinical testing relies on targeted, kit-based RT-PCR panels. These are limited to a predefined set of known pathogens (e.g., Influenza, RSV, SARS-CoV-2) and often miss low-abundance infections, novel organisms, or unexpected co-infections.
• Clinical Gap: In New Zealand, where respiratory hospitalization rates are high and disproportionately affect Māori and Pacific peoples, a substantial fraction of SARI cases remain without a confirmed cause, hindering appropriate clinical management and public health surveillance.

2. Methodology

The study employed an unbiased metatranscriptomic sequencing approach to profile the total "infectome" (the community of actively transcribing pathogens) in patients who tested negative via standard PCR.

• Cohort: 300 nasopharyngeal swab samples were selected from a larger pool of 1,274 SARI patients collected between 2014 and 2021 through the SHIVERS surveillance program in Auckland, New Zealand. All 300 samples were confirmed PCR-negative for a standard panel (Influenza A/B, Adenovirus, RSV, hMPV, Parainfluenza 1-3, Rhinovirus, Enterovirus, and SARS-CoV-2 from 2020).
• Sequencing Workflow:
  • Extraction: Total RNA was re-extracted and ribosomal RNA (rRNA) was depleted using the Illumina Stranded Total RNA Prep with Ribo-Zero Plus kit.
  • Sequencing: Paired-end (150 bp) sequencing was performed on the Illumina NovaSeqX platform, generating an average of 50.2 million raw reads per library.
• Bioinformatics Pipeline:
  • Preprocessing: Quality trimming (Trimmomatic) and removal of human host reads (Bowtie2 mapping to GRCh38).
  • Assembly: De novo assembly of remaining reads into contigs using MEGAHIT.
  • Taxonomic Classification:
    • Viruses: BLASTx/BLASTn against NCBI databases; phylogenetic analysis using IQ-TREE; specific typing for Rhinovirus via VP1 gene.
    • Non-Viral: CCMetagen (using KMA alignment) for species-level classification of bacteria and fungi.
  • Functional Analysis:
    • Virulence: Mapping reads to reference genomes of top bacterial species to quantify expression of virulence factors (VFDB database).
    • Resistome: Mapping reads to the ResFinder database to identify Antimicrobial Resistance Genes (ARGs).
• Statistical Analysis: Assessment of alpha/beta diversity (Shannon, Simpson, Bray-Curtis), PERMANOVA for community clustering, and network analysis for co-infection patterns.

3. Key Contributions

• First Metatranscriptomic Characterization: This is the first study to apply metatranscriptomics to undiagnosed SARI cases in New Zealand, providing a comprehensive view of the respiratory microbiome.
• Functional Profiling: Unlike metagenomics (which detects DNA presence), metatranscriptomics captured actively transcribing organisms, offering insights into which pathogens were metabolically active and expressing virulence or resistance genes.
• Diagnostic Validation: The study demonstrated that standard PCR panels can yield false negatives even for common pathogens (like Rhinovirus) due to primer mismatches or low abundance, highlighting the need for broader surveillance.

4. Key Results

A. Pathogen Discovery

• Detection Rate: 43% (129/299) of PCR-negative samples contained actively transcribing potential pathogens.
• Diversity: The study identified:
  • 10 RNA Viruses: Including Human Rhinovirus A, B, and C (most frequent), Human Coronaviruses (229E, OC43, HKU1, NL63), Measles virus (Genotype B3), Parainfluenza virus type 4, and Human Parechovirus A1.
  • 3 DNA Viruses: Human alphaherpesvirus 3 (Varicella-zoster), Human gammaherpesvirus 4 (EBV), and Papillomaviruses (including a feline strain).
  • 9 Bacterial Species: Dominated by Streptococcus pneumoniae, Pseudomonas aeruginosa, Staphylococcus aureus, and Moraxella catarrhalis.
  • 4 Fungal Species: Including Naganishia albida (a rare opportunistic pathogen) and Penicillium chrysogenum.

B. Co-infections

• Prevalence: 26% of the positive cases (33/129) involved co-infections with two or more microbial species.
• Patterns: Common co-infections included Rhinovirus C with Moraxella catarrhalis, and S. pneumoniae with S. aureus.
• Complexity: Some samples contained up to five distinct microbial species, suggesting polymicrobial interactions that standard diagnostics miss.

C. Virulence and Resistance

• Virulence Factors: Highly expressed virulence genes were detected in dominant bacteria, including OmpA (adhesion) in M. catarrhalis, SdrD and ClfB (adhesion/colonization) in S. aureus, and PrtA (protease) in S. pneumoniae.
• Antimicrobial Resistance (ARGs): 438 ARGs were identified. Notable findings included the tmexD1 efflux pump gene (conferring multi-class resistance) in Klebsiella pneumoniae and blaZ (beta-lactamase) in S. aureus. Notably, the mecA gene (methicillin resistance) was not detected.

D. Epidemiological Drivers

• Demographics: Age and year of collection were the strongest predictors of infectome composition and diversity.
  • Age: Children (0–5 years) had the highest proportion of viral detections. The >65 age group showed the highest species richness.
  • Temporal: Significant variation was observed between years (e.g., 2015 vs. 2016), likely reflecting seasonal pathogen dynamics.
• PCR Failure: Alignment of detected Rhinovirus sequences against the PCR primers used in routine testing showed perfect complementarity, suggesting the PCR failures were likely due to low viral load or stochastic sampling issues rather than primer mismatch.

5. Significance and Implications

• Clinical Impact: The study reveals that a large proportion of "undiagnosed" SARI cases are actually caused by known pathogens missed by routine panels, or by complex polymicrobial infections. This suggests that current diagnostic strategies may lead to delayed treatment or inappropriate antibiotic use.
• Public Health Surveillance: Integrating metatranscriptomics into surveillance could improve the detection of emerging pathogens (e.g., Measles, rare coronaviruses) and track the evolution of antimicrobial resistance in real-time.
• Future Directions: The authors advocate for the development of genomic-informed diagnostic pipelines in New Zealand to complement existing PCR workflows. This would allow for more accurate etiological diagnosis, better understanding of co-infection dynamics, and enhanced preparedness for future outbreaks.

Conclusion: Metatranscriptomics provides a powerful, unbiased tool to uncover the "hidden" infectome in SARI, revealing a complex landscape of viral, bacterial, and fungal co-infections that drive disease severity and are invisible to conventional diagnostics.