Leveraging 16S rRNA data to uncover vaginal microbial signatures in women with cervical cancer.

ABSTRACT: Microbiota-relevant signatures have been investigated for human papillomavirus-related cervical cancer (CC), but lack consistency because of study- and methodology-derived heterogeneities. Here, four publicly available 16S rRNA datasets including 171 vaginal samples (51 CC versus 120 healthy controls) were analyzed to characterize reproducible CC-associated microbial signatures. We employed a recently published clustering approach called VAginaL community state typE Nearest CentroId clAssifier to assign the metadata to 13 community state types (CSTs) in our study. Nine subCSTs were identified. A random forest model (RFM) classifier was constructed to identify 33 optimal genus-based and 94 species-based signatures. Confounder analysis revealed confounding effects on both study- and hypervar

SUBMITTER: Wu M 

PROVIDER: S-EPMC9892946 | biostudies-literature | 2023

REPOSITORIES: biostudies-literature