Identification of pleural infection microbiological patterns by applying next generation sequencing and bioinformatics analysis

Background: Pleural infection (PI) is a common and complex disease which can be life threatening for immunocompromised and elderly populations. Prior antibiotic use and special bacterial nutritional requirements hamper the accuracy of bacterial identification using current clinical culture-based techniques. Consequently, PI microbiology remains unclear. Next generation sequencing (NGS) has the potential to improve identification of the total bacterial population of a complex sample. Aim: To discover and characterise the microbial patterns of PI using NGS and bioinformatics techniques. Methods: Pleural fluid samples from the “Pleural Infection Longitudinal Outcome Study” (PILOT, ISRCTN50236700, n=243) underwent bacterial DNA extraction followed by 16S rRNA NGS using Illumina MiSeq. Data were analysed with DADA2 and Phyloseq R packages. Results: Analysis showed diverse microbiological patterns for PI as 391 different pathogens were identified up to the genus level. 131 (54%) samples had one pathogen with relative abundance over 50% and 89 (36%) samples had at least three pathogens with relative abundance over 10%, suggesting a polymicrobial infection. Streptococcus pneumoniae was detected in 40 (16%) and Staphylococcus aureus in 20 (8%) samples. Discussion: We established a methodology to extract bacterial DNA from patients with PI and used it as a template to apply NGS. 16S rRNA gene NGS provides a robust method to investigate the bacteriological patterns in pleural fluid of patients with PI. Funding: National Institute for Health Research, Oxford Biomedical Research Centre