Aim
The aim of this doctoral dissertation was to investigate the clinical utility of metagenomic sequencing (mNGS) in the diagnosis of patients with acute undifferentiated fever and no obvious localization of infection in Slovenia. We examined to what extent mNGS enables pathogen identification in patients in the acute phase of the disease and developed protocols for incorporating mNGS into clinical diagnostics. For this purpose, we developed a comprehensive mNGS protocol and defined metrics for standardized reporting of mNGS results.
Hypotheses
(H1) Using metagenomic sequencing, we will identify the causative pathogen in 60–80 % of patients with a previously microbiologically confirmed cause of acute undifferentiated fever without clear signs of organ involvement. (H2) Using metagenomic sequencing, we will identify the pathogen in 15–25 % of patients with etiologically unexplained acute undifferentiated fever without clear signs of organ involvement. (H3) Using metagenomic sequencing, we will detect emerging viral causes of fever that present without clear signs of organ involvement in Slovenia. (H4) We will develop a more comprehensive metagenomic sequencing workflow capable of efficiently detecting infections caused by RNA viruses.
Methods
The doctoral dissertation included whole-blood samples stored from 462 patients. Sequencing was performed using both short-read and long-read technologies. The doctoral dissertation is composed of three parts. In the first part, we examined the impact of modifications in individual sample preparation steps on sequencing results and data analysis. In the second part, we evaluated the optimized protocol using clinical samples from patients in whom the pathogen had previously been confirmed in acute blood samples using established molecular microbiological methods. In the third part, we applied the evaluated mNGS protocol to search for microorganisms in blood samples from patients with acute undifferentiated fever and no obvious localization of infection whose etiology remained unidentified using routine diagnostic approaches.
Results
Based on the results of the first part of the doctoral dissertation, we developed an innovative approach for sequencing blood samples from patients with acute acute undifferentiated fever and no obvious localization of infection, enabling detection of both intracellular and extracellular microorganisms with DNA and RNA genomes. This approach preserves detection of DNA pathogens while enhancing RNA yield and reducing host nucleic acid. The protocol combines SISPA-A-prepared nucleic acid isolates from whole blood with SISPA-B-prepared isolates from plasma, further treated with DNase and QIAseq FastSelect rRNA/Globin reagents. In the second part of the doctoral dissertation, we included 200 patients. Among them, there were 97 infections caused by DNA pathogens and 103 infections caused by RNA pathogens. Etiologically, there were 79 bacterial, three parasitic, and 118 viral infections. Among viral pathogens, 15 samples contained DNA viruses and 103 contained RNA viruses. By manual review of sequencing results, mNGS detected the same pathogen in 79.5 % of cases (159/200) where the pathogen had been previously confirmed by molecular methods. Comparison of the ClinSeq tool with manual analysis showed 83.5 % overall agreement (OPA), 79.2 % positive percent agreement (PPA), and 100 % negative percent agreement (NPA). The Cohen’s kappa (κ) was 0.61, indicating substantial agreement between the methods. In the final part, we analyzed 262 patient samples. Among 53 patients with serologically confirmed etiology, mNGS detected the pathogen in the acute sample in 5.7 % of cases (3/53). Among 209 patients with acute undifferentiated fever and no obvious localization of infection, etiology was determined in 11.0 % of cases (23/209), a possible pathogen was identified in 2.4 % (5/209), while 86.6 % (181/209) remained unexplained. Using mNGS, we also detected infections with Borrelia miyamotoi and Spiroplasma ixodetis for the first time.
Conclusions
The developed mNGS workflow enables reliable detection of a wide range of pathogens, including non-cultivable bacteria, viruses, and parasites. Combining plasma and whole-blood samples prior to library preparation improves detection, simplifies processing, reduces costs, and allows flexible sequencing options. Introducing a mathematical classification method further enhances result interpretation. These findings represent the first confirmed cases of B. miyamotoi and S. ixodetis infections in Slovenia and provide a foundation for incorporating mNGS into routine diagnostics. Future studies will focus on validating the protocol on additional sample types, comparing it with other routine methods, and optimizing analytical approaches for maximal clinical utility.
Key Findings
(1) For patients with acute undifferentiated fever and no obvious localization of infection, mNGS sensitivity was 79.5 %. (2) mNGS identified the pathogen in additional 11.0 % of patients with etiologically unexplained fever. (3) mNGS confirmed the first cases of B. miyamotoi and S. ixodetis infections in Slovenia. (4) mNGS detected a co-infection with Neoehrlichia mikurensis in a patient with hemorrhagic fever and renal syndrome. (5) We developed a comprehensive protocol for metagenomic sequencing of blood samples from patients with acute undifferentiated fever and no obvious localization of infection. (6) We developed the standardized bioinformatics pipeline MetaAll and the ClinSeq tool, which are freely available and allow efficient pathogen detection in clinical samples. (7) Using ClinSeq, we standardized reporting of metagenomic results and defined criteria for mNGS interpretation.
|
