A comparison of two culture and three sequencing methods for detection of the urobiome

Abstract

Standard urine culture (SUC), the clinical reference method, is optimized for fast-growing aerobes and misses much of the urinary microbiota. Enhanced culture and sequencing methods have rarely been compared on the same specimens. We compared five methods (SUC, expanded quantitative urine culture [EQUC], short-read V4 16S rRNA gene sequencing, long-read near full-length 16S rRNA gene sequencing, and PCR-free long-read shotgun metagenomic sequencing) on parallel aliquots of urine from pregnant participants screened for asymptomatic bacteriuria. Of 98 cultured samples, 98 (100.0%) were EQUC-positive while 85 (86.7%) SUC-positive. EQUC detected 102 species, but SUC only detected 41. Short-read 16S rRNA gene sequencing, long-read 16S rRNA gene sequencing, and shotgun sequencing were positive for 188/188, 186/188 and 139/139 samples, detecting 195, 69 and 528 species, respectively. In the 79 samples positive by EQUC and all three sequencing methods, sequencing profiles resembled each other more than culture profiles, and SUC captured only a subset of taxa detected by every other method. Shotgun sequencing gave the broadest taxonomic coverage and was the only sequencing method to identify Escherichia coli to the species level. With EQUC, it outperformed both 16S sequencing methods in detecting Gardnerella vaginalis and distinguishing Lactobacillus species, and it resolved six Gardnerella species, three detected by no other method. Long-read 16S rRNA gene sequencing gave the most confident amplicon-based species assignments, whereas culture better recovered Corynebacterium and Staphylococcus species. No single method was best across all endpoints, but long-read shotgun metagenomics provided the broadest species-level view. Method choice should match the research or clinical question; outcome-based studies are needed to establish clinical utility.