RESUMO
The microbiome from the reproductive tract is being investigated for its putative effect on fertility, embryo development, and health status of the human or animal host postpartum. Besides the presence of a vaginal microbiome, recent studies have claimed the existence and putative role of the uterine microbiome. Yet, the extremely low bacterial numbers and high eukaryotic/prokaryotic DNA ratio make this a highly challenging environment to study with next-generation sequencing (NGS) techniques. Here, we describe the methodological challenges that are typically encountered when performing an accurate analysis of low microbial biomass samples, illustrated by data of our own observational study. In terms of the research question, we compared the microbial composition throughout different parts of the reproductive tract of clinically healthy, mid-lactation Holstein-Friesian cows. Samples were collected from 5 dairy cows immediately after killing. Swabs were taken from the vagina, and from 4 pre-established locations of the uterine endometrium. In addition to the conventional DNA extraction blank controls, sterile swabs rubbed over disinfected disposable gloves and the disinfected surface of the uterus (tunica serosa) before incision were taken as sampling controls. The DNA extraction, DNA quantification, quantitative PCR of the 16S rRNA genes, and 16S rRNA gene sequencing were performed. In terms of NGS data analysis, we performed prevalence-based filtering of putative contaminant operational taxonomic units (OTU) using the decontam R package. Although the bacterial composition differed between the vagina and uterus, no differences in bacterial community structure (α and ß diversity) were found among the different locations in the uterus. At phylum level, uterine samples had a greater relative abundance of Proteobacteria, and a lesser relative abundance of Firmicutes than vaginal samples. The number of shared OTU between vagina and uterus was limited, suggesting the existence of bacterial transmission routes other than the transcervical one to the uterus. The mid-lactation bovine genital tract is a low microbial biomass environment, which makes it difficult to distinguish between its constitutive versus contaminant microbiome. The integration of key controls is therefore strictly necessary to decrease the effect of accidentally introduced contaminant sequences and improve the reliability of results in samples with low microbial biomass.