TL;DR
This study explores the use of metagenomic sequencing (NGS and Nanopore) to identify pathogens and their antibiotic resistance profiles in culture-negative infective endocarditis (IE) cases. The findings suggest that these advanced sequencing techniques can reliably detect causative agents and provide crucial information for guiding treatment, offering a powerful alternative to traditional culture methods.
What Problem Does This Paper Address?
Infective endocarditis (IE) is a serious heart valve infection that can lead to significant morbidity and mortality. Accurate diagnosis and timely treatment are crucial for patient outcomes. The gold standard for identifying the causative pathogens has been blood and valve cultures. However, a significant challenge arises when these cultures are negative, a common occurrence in patients who have already received antibiotic treatment. These "culture-negative" cases make precise diagnosis difficult, hindering effective treatment. Traditional methods like 16S rRNA sequencing can identify bacteria but lack information on antibiotic resistance. Metagenomic sequencing offers a promising solution by directly analyzing all genetic material in a sample, potentially identifying a wide range of pathogens (bacteria, viruses, fungi) and their antibiotic resistance genes without the need for culturing. This paper specifically investigates the feasibility and effectiveness of using both Next-Generation Sequencing (NGS) and Oxford Nanopore Technologies MinION sequencing for this purpose in culture-negative IE patients.
Methods
The researchers employed two primary sequencing technologies:
- Next-Generation Sequencing (NGS): DNA extracted from resected heart valves was fragmented, libraries were prepared, and sequencing was performed on a BGISEQ-500 platform. This method generates short reads.
- Oxford Nanopore Technologies MinION: DNA was also prepared for sequencing using the MinION device, which produces long reads.
Both datasets were processed using an in-house developed bioinformatics pipeline. Key steps included:
- Quality Control: Removing low-quality reads, adapter contamination, and duplicated reads.
- Host Depletion: Aligning reads to the human genome to remove host DNA and retain microbial sequences.
- Species Identification: Using the Centrifuge k-mer alignment algorithm and BLAST searches against microbial genome databases to identify pathogens.
- Antimicrobial Resistance (AMR) Analysis: Aligning reads to the CARD (Comprehensive Antibiotic Resistance Database) using BLASTn to identify AMR genes and single nucleotide polymorphisms (SNPs).
Sanger sequencing was used for validation of the metagenomic findings.
Datasets
The study utilized resected valve tissue samples from seven clinically diagnosed, culture-negative IE patients. No new benchmark datasets were introduced; the analysis relied on publicly available microbial genome databases from NCBI.
Key Insights and Ideas
- Metagenomics as a powerful tool for culture-negative IE: The study strongly supports metagenomic sequencing as a viable and effective method for identifying pathogens in IE cases where traditional cultures fail.
- Complementary strengths of NGS and Nanopore: While both platforms could identify pathogens, NGS provided broader coverage and detected more AMR features, likely due to its higher data output. Nanopore sequencing, with its long reads, offered higher specificity in species identification, particularly for closely related bacteria.
- Real-time potential of Nanopore: The MinION platform demonstrated the potential for rapid, real-time pathogen detection, with results potentially available within hours, which is crucial for timely clinical decision-making.
- Challenges in AMR detection: Accurately linking AMR genes to specific microbes remains a challenge, especially with highly homologous genes across different species. The study highlights the need for comprehensive AMR databases and potentially clinic-specific libraries.
- Host DNA contamination is a major hurdle: A significant portion of sequencing data (over 95%) was mapped to the human genome, representing a substantial waste of sequencing resources and impacting detection sensitivity.
Key Results
- Both NGS and Nanopore sequencing reliably identified causative pathogens in all seven samples, largely in agreement with Sanger sequencing (with one exception where Sanger failed).
- NGS and Nanopore sequencing could identify bacteria to the species level, a finer resolution than Sanger sequencing in some cases.
- NGS identified a significantly higher number of different bacteria per sample compared to Nanopore sequencing, attributed to its higher data output.
- More AMR features were detected using NGS due to better genome coverage.
- Nanopore sequencing showed promise for rapid detection, with pathogens identified within 1-4 hours depending on the read cutoff.
Limitations and Future Work
- Sample multiplexing in Nanopore: The study noted that barcode leaking (sample cross-contamination) was more apparent with Nanopore sequencing due to its lower base call accuracy. They recommend sequencing only one sample per flow cell for clinical settings to avoid contamination and improve turnaround time.
- Host DNA depletion: The high percentage of host genome reads is a significant limitation. The authors suggest that developing appropriate host depletion methods before library preparation is critical to improve the efficiency and sensitivity of pathogen detection.
- AMR gene origin: While NGS detected more AMR features, it struggled to pinpoint the exact bacterial origin of these genes. Future work could focus on developing methods to better link AMR genes to specific microbes, potentially leveraging the long-read capabilities of Nanopore sequencing.
- Clinic-specific AMR libraries: The authors suggest the potential need for constructing clinic-specific AMR libraries to enhance the detection of AMR features relevant to local pathogens.
Further studies are recommended to optimize sample processing, sequencing protocols, and bioinformatics analysis for clinical microorganism detection using metagenomics.
Summary updated