Rapid diagnosis of Mycoplasma pneumoniae and prediction of antibiotic resistance by nanopore adaptive sampling

Abstract

Background

Microbial infections, particularly in children, require rapid and accurate diagnostics. It is difficult to differentiate pathogens from commensal organisms, and it is impossible to identify antibiotic resistance genes that belong to pathogens with current methods. Third-generation sequencing provides rapid library preparation and real-time data acquisition. Nanopore normal sampling (NNS) enables unbiased sequencing of clinical samples without amplification, aiding pathogen identification and antimicrobial resistance gene prediction. However, clinical samples often contain a considerable amount of human DNA, potentially masking pathogen data. Nanopore adaptive sampling (NAS) aims to selectively enrich pathogens, promising improved diagnostics for acute infections and better treatment decisions in clinical practice. This study aimed to determine the utility of NAS in enhancing the real-time detection of pathogens and predicting AMR in infectious disease outbreaks.

Methods

This study used NAS technology to rapidly and directly detect Mycoplasma pneumoniae infection in bronchoalveolar lavage fluid samples from 28 pediatric patients at Shenzhen Children's Hospital. We assessed the efficacy of NAS compared with that of NNS by evaluating the number of microbial reads and the amount of microbial DNA data. We then compared the accuracy of detecting pathogens between NNS and NAS and between NAS and real-time polymerase chain reaction assays. Furthermore, we predicted antimicrobial resistance (AMR) and examined AMR genes associated with pathogens.

Results

NAS showed up to a 14.67-fold increase in the amount of microbial DNA data from patients' samples compared with NNS within the initial 2.5 h of sequencing. Additionally, NAS reduced the amount of host DNA data by up to 6.67-fold compared with NNS. Unlike TaqMan real-time polymerase chain reaction assays, NAS technology identified dominant pathogens and provided detailed insight into the abundance of the microbial community. Furthermore, NAS was able to predict AMR profiles of microbial communities and attribute specific AMR traits to individual microbes within the samples.

Conclusion

This study shows that NAS advances the clinical diagnosis because it can rapidly detect pathogens directly from patients' samples and provides antimicrobial resistance information for clinical guidance. These abilities further facilitate the application of NAS in personalized treatment, reduce the misuse of broad-spectrum antibiotics, and promote patients' recovery.