Neovaginas are surgically constructed to address uterovaginal agenesis in Mayer-Rokitansky-Küster-Hauser (MRKH) syndrome or as part of gender-affirming care for transfeminine individuals. Prior research, often cross-sectional and limited by methodology, has shown a polymicrobial environment in neovaginas, particularly in transgender women. A recent metaproteomic study highlighted differences between the microbiota of penile skin-lined neovaginas and normal vaginas. The host epithelium plays a crucial role in shaping the vaginal microbiome. In healthy individuals, the stratified squamous epithelium, coated with cervicovaginal mucus, provides a habitat and nutrients for microbes. During reproductive years, high glycogen levels favor *Lactobacillus* dominance. In contrast, reduced estrogen post-menopause lowers glycogen, leading to reduced *Lactobacillus* and an increase in BV-associated anaerobes. While studies suggest neovaginas develop a stratified squamous epithelium over time, detailed information on neovaginal microbiota assembly is lacking. This study uniquely follows the development of a vaginal microbiota in surgically constructed neovaginas, offering insights into this process.

Previous studies on neovaginal microbiota have been limited, primarily using cultivation and microscopy. These studies, often focusing on transgender individuals, generally identified a polymicrobial environment. One study using metaproteomics demonstrated distinct differences between the microbiota of penile skin-lined neovaginas and those of typical vaginas. Existing literature highlights the importance of the host epithelium in shaping the vaginal microbiome, with glycogen levels influencing *Lactobacillus* dominance. The transition from a peritoneal lining to a stratified squamous epithelium in surgically created neovaginas has been noted, but comprehensive characterization of the associated microbial community development is lacking. The assembly of the human gut microbiota in infants has been extensively studied, offering a useful framework for comparison, although the establishment of the vaginal microbiota is less understood. The unique opportunity provided by surgically constructed neovaginas in MRKH patients allows for a detailed examination of vaginal microbiota establishment.

This longitudinal study enrolled 39 Chinese women with MRKH syndrome (age 17-35 years). Laparoscopic peritoneal vaginoplasty was performed. Multi-site longitudinal samples were collected before surgery (PRE) and at four post-surgical time points (P14D, P90D, P6/12M, P2/4Y). A total of 456 samples were collected and analyzed using deep shotgun metagenomic sequencing. These included vaginal, peritoneal fluid, fecal, tongue coating, abdominal skin, and control saline samples. Metadata were gathered through clinical examination and questionnaires. For comparison, data from 472 healthy adult Chinese women (pre- and postmenopausal) were used as references. Deep shotgun metagenomic sequencing was conducted, and data were preprocessed using various bioinformatics tools. Microbiota alpha diversity was assessed, and Bray-Curtis dissimilarity was calculated. Dirichlet Multinomial Mixtures (DMM) were used for community typing. Co-occurrence networks were analyzed using SparCC. Source tracking was performed using FEAST and StrainPhlAn to identify strain transmission. Generalized estimating equations (GEE) examined the impact of pre-surgery microbiota on neovaginal microbiota development. A random forest model was used to predict temporal changes in neovaginal microbiota abundance.

Pre-surgery dimple microbiota composition, including species like *Lactobacillus iners*, *Prevotella timonensis*, *Atopobium vaginae*, showed resemblance to both pre- and postmenopausal women. At P14D, species numbers and relative abundances of certain species decreased, while others, such as *Enterococcus faecalis* and genital Mycoplasmas, bloomed due to antibiotic resistance. Despite drastic changes at P14D, intra-individual microbiota dissimilarity was lower than inter-individual dissimilarity. At P90D, the microbiota developed into a polymicrobial state with increased BV-associated species. At P2/4Y, the neovaginal microbiota closely resembled the pre-surgery dimple microbiota, indicating restoration of homeostasis. DMM analysis identified three clusters, with a biased transition towards a *Prevotella timonensis* dominated cluster at P2/4Y. Source tracking revealed the pre-surgery dimple microbiota as the major contributor to neovaginal colonization, followed by fecal microbiota. Strain-level analysis confirmed this, with *Lactobacillus crispatus* showing primary origin from the dimple. MRKH type correlated with the relative abundance of specific species. Pre-surgery dimple microbiota status significantly correlated with the relative abundance of several species in the neovagina, including *Lactobacillus crispatus*. Random forest models effectively predicted temporal changes in the relative abundance of certain neovaginal species.

The findings demonstrate a dynamic interplay between initial microbial seeding, host-driven environmental changes (epithelial development and glycogen production), and selective colonization in shaping the neovaginal microbiota. The early stochastic nature gives way to a more structured community resembling a normal vagina, albeit initially with BV-like characteristics. The eventual return to a microbiota composition similar to the pre-surgery dimple suggests the importance of this initial inoculum and the influence of the evolving neovaginal environment. The high prevalence and persistence of *Lactobacillus crispatus* from the dimple highlight its potential role in establishing a healthy neovaginal microbiota, and points to future research opportunities on promoting its colonization. This study parallels findings in infant gut microbiota assembly where early perturbations recover over time. The different microbiota observed in penile skin-lined neovaginas highlights the influence of the tissue source in neovaginal microbiota establishment. Ecological network analysis showed the development of a more complex and stable network at later timepoints. The identification of keystone taxa like *Prevotella timonensis* may pave the way for microbial ecology engineering.

This study provides novel insights into the temporal dynamics of neovaginal microbiota assembly in MRKH patients, revealing a complex interplay between the initial microbial inoculum, the host's evolving environment, and selective colonization. The findings underscore the potential of targeting specific species, such as *Lactobacillus crispatus*, to promote a healthy neovaginal microbiota. Further research should explore the impact of various factors influencing microbiota assembly, including peri-surgical interventions and individualized responses, to optimize neovaginal health following vaginoplasty. Investigating the role of keystone species in driving microbiota assembly holds potential for developing targeted interventions.

The study's sample size, while substantial, could be expanded for greater statistical power. The COVID-19 pandemic impacted sample collection at the 6-12 months time point, potentially introducing bias. Individual responses to peri-surgical medications and other factors could not be fully accounted for. The findings are specific to MRKH patients and might not be directly generalizable to other populations undergoing neovaginoplasty.