Development and application of clinical metagenomics for the diagnosis and characterisation of lower respiratory infections.

Lower respiratory tract infections (LRTIs) are the largest infectious cause of death globally according to the WHO. Rapid diagnosis and appropriate management are key to control. The gold standard for diagnosis, microbiological culture, is too slow to provide clinicians with the necessary information to treat patients appropriately and reduce morbidity and mortality. In recent years, there has been growing interest in the application of clinical metagenomics (CMg) for the characterisation of pathogens in clinical samples. CMg has the potential to be faster and more comprehensive than culture, capable of detecting any pathogen (bacteria, viruses and fungi) in a single test within hours. This would transform the field of clinical microbiology, ensuring patients received the appropriate antibiotics sooner, while concurrently providing pathogen genomic surveillance data. However, the development and implementation of rapid CMg has been challenging, mainly due to the high ratio of human:pathogen DNA present in clinical samples, resulting in high sequencing cost and long turnaround times. In this study, we developed and optimised a rapid CMg pipeline for the diagnosis and characterisation of LRTIs that overcomes these challenges. The pipeline includes: a simple and highly efficient saponin-based host DNA depletion step, automated microbial DNA extraction, rapid library preparation, low-input nanopore sequencing and real-time identification of microorganisms and resistance genes. The pipeline was developed, tested and optimised using excess respiratory samples from suspected LRTI patients. The optimised pipeline was then evaluated in a clinical trial comparing three technologies for the rapid diagnosis of hospital acquired (HAP) and ventilator associated (VAP) pneumonia (the INHALE trial). The pipeline was also evaluated for the rapid characterisation of Legionella spp. in respiratory samples for outbreak investigation applications in collaboration with Public Health England. The developed CMg test had a turnaround time of six hours for the identification of bacterial pathogens and resistance genes with high specificity and sensitivity. A reduction in sensitivity was observed when applying the method for the detection of pathogens in suspected HAP/VAP patients and for samples containing Legionella spp. Reduced performance was related to the difference in how respiratory samples from the intensive care unit are cultured and testing of old freeze thawed samples, respectively. CMg demonstrates great potential for replacing culture for the diagnosis of LRTIs, however, further optimisation is required to enable concurrent detection of viruses and improved automation is required to allow successful clinical implementation