Abstract
Despite a substantial increase in the demand for assisted reproductive technologies (ART)—including in vitro fertilisation (IVF) and intracytoplasmic sperm injection (ICSI)—successful outcomes from these treatments have not correspondingly increased. Many factors impact ART success, with infertility remaining significant. While overt causes of infertility have been widely investigated, emerging research has begun to explore underlying factors, such as the interaction between host and microbial communities. However, the relationship between the vaginal microbiome and fertility remains poorly understood, with methodologies for accurately characterising this microbiome still evolving. Current approaches rely on short-read DNA sequencing of 16S rRNA gene fragments of bacteria, hindering species-specific identification. Nanopore sequencing has emerged as a promising approach, providing coverage of the entire 16S rRNA gene through large-fragment sequencing. Despite this, the lack of standardisation and validation, along with challenges in primer design, impacts microbial accuracy identification. This study aimed to investigate the utility of whole 16S sequencing for characterisation of the human vaginal microbiome via the development of a reproducible ONT sequencing assay for microbial detection.To address this, four sub-studies were performed beginning with a review of the current methodological approaches used in assessing the vaginal microbiome of the female reproductive tract when undergoing fertility treatment. This provided an understanding of the most common methods as well as the strengths and limitations of these approaches when assessing the vaginal microbiome. Key findings found a lack of uniformity in sample collection and transport, DNA extraction, 16S amplification vs a whole genome approach, next-generation sequencing (NGS) library preparation, and bioinformatic analyses. Nanopore sequencing was identified as a promising tool for full-length 16S rRNA investigations, with further studies needed to optimise and standardise relevant methods. Utilising this information, the next phase of the approach focused on the development and refinement of the relevant methodologies including primer design, PCR optimisation, and bioinformatic components of the Nanopore sequencing protocol. During this, an additional focus was placed on Chlamydia trachomatis spiking to assess primer accuracy using an oral microbiome sample set. Improved PCR amplification was achieved using degenerative primers at the 16S rRNA region for 27F-YM and 341F-NW primers, consistent with previous literature. However, all primer pairs failed in the detection of C. trachomatis within the oral microbiome samples, highlighting the need for further refinement and reflects current challenges associated with primers for 16S rRNA sequencing of the vaginal microbiome.
The investigation then focused on the development and comparison of analysis tools including bioinformatic pipelines and nanopore-compatible tools to evaluate the accuracy of microbial detection and benchmarking against publicly available data [PRJEB53337]. Investigation of bioinformatic approaches identified Porechop and NanoCLUST as the most accurate tools in identifying microbial communities when benchmarked against publicly available data. The final stage of this investigation involved a pilot study of the experimental and analysis methodologies to assess the most abundant vaginal microbes in a healthy population (n=12) of females. Lactobacillus dominance was observed in all vaginal microbiome samples with community state type (CST) I as the most common vaginal classification which reflects advantageous reproductive health outcomes demonstrated throughout literature. However, this work found slight discrepancies in Lactobacillus species between primers 341F-NW and 27F-YM_MIX across vaginal samples, possibly due to variations in read depth, indicating that there may be sample location-specific biases that need to be accounted for in microbiome investigations. Read depth within oral or real vaginal samples varied significantly between samples possibly influencing microbial detection. Additionally, samples collected throughout this study—oral or vaginal samples—may have contained large amounts of host DNA which was not removed and may have reduced the overall resolution of microbial DNA. Stringent bioinformatic analysis required a high computational load which was unable to be met leading to less stringent parameters and possibly influencing microbial detection and representation.
Overall, sequencing of both oral and true microbiome samples underscored the need for further validation to enable accurate microbial identification. A significant need for data deposition and availability to allow method benchmarking and standardisation was highlighted for future studies. A whole genome approach with adaptive sampling may enhance microbial representation and detect organisms such as fungi, micro-eukaryotes, and viruses. This would also overcome limited C. trachomatis detection in 16S rRNA studies due to minor primer nucleotide mismatches. If 16S rRNA sequencing is chosen, future studies should focus on PCR optimisation and primer re-design to accurately detect microbes such as C. trachomatis. While this thesis has made strides in addressing methodological gaps in microbial investigations, further research is essential to refine sequencing methodologies for the vaginal microbiome and to explore the role of microbial interactions in fertility outcomes. Overcoming the methodological gaps in microbial investigations is vital as it will aid clinicians in accurately characterising microbial communities and enable larger streamlined studies to further analyse the impact of the vaginal microbiome on fertility outcomes. Resolving common pitfalls in methodologies will also allow future investigations to assess the suitability of Nanopore sequencing as a potential diagnostic tool for clinicians to accurately characterise a microbiome in a cost and time-efficient manner.
| Date of Award | 2025 |
|---|---|
| Original language | English |
| Supervisor | Paul Dunn (Supervisor), Larisa M Haupt (Supervisor), Elham Nikbakht Nasrabadi (Supervisor) & Kevin Ashton (Supervisor) |