Biology subjects

Brown Kav, A.

Publications and source records attributed to Brown Kav, A..

3 recordsLinked to original sources

bioRxiv2024

Early prediction of preeclampsia using the first trimester vaginal microbiome

Preeclampsia is a severe obstetrical syndrome which contributes to 10-15% of all maternal deaths. Although the mechanisms underlying systemic damage in preeclampsia--such as impaired placentation, endothelial dysfunction, and immune dysregulation--are well studied, the initial triggers of the condition remain largely unknown. Furthermore, although the pathogenesis of preeclampsia begins early in pregnancy, there are no early diagnostics for this life-threatening syndrome, which is typically diagnosed much later, after systemic damage has already manifested. Here, we performed deep metagenomic sequencing and multiplex immunoassays of vaginal samples collected during the first trimester from 124 pregnant individuals, including 62 who developed preeclampsia with severe features. We identified multiple significant associations between vaginal immune factors, microbes, clinical factors, and the early pathogenesis of preeclampsia. These associations vary with BMI, and stratification revealed strong associations between preeclampsia and Bifidobacterium spp., Prevotella timonensis, and Sneathia vaginalis. Finally, we developed machine learning models that predict the development of preeclampsia using this first trimester data, collected ~5.7 months prior to clinical diagnosis, with an auROC of 0.78. We validated our models using data from an independent cohort (MOMS-PI), achieving an auROC of 0.80. Our findings highlight robust associations among the vaginal microbiome, local host immunity, and early pathogenic processes of preeclampsia, paving the way for early detection, prevention and intervention for this devastating condition.

microbiology↗

bioRxiv2024

Processing-bias correction with DEBIAS-M improves cross-study generalization of microbiome-based prediction models

Every step in common microbiome profiling protocols has variable efficiency for each microbe. For example, different DNA extraction kits may have different efficiency for Gram-positive and -negative bacteria. These variable efficiencies, combined with technical variation, create strong processing biases, which impede the identification of signals that are reproducible across studies and the development of generalizable and biologically interpretable prediction models. "Batch-correction" methods have been used to alleviate these issues computationally with some success. However, many make strong parametric assumptions which do not necessarily apply to microbiome data or processing biases, or require the use of an outcome variable, which risks overfitting. Lastly and importantly, existing transformations used to correct microbiome data are largely non-interpretable, and could, for example, introduce values to features that were initially mostly zeros. Altogether, processing bias currently compromises our ability to glean robust and generalizable biological insights from microbiome data. Here, we present DEBIAS-M (Domain adaptation with phenotype Estimation and Batch Integration Across Studies of the Microbiome), an interpretable framework for inference and correction of processing bias, which facilitates domain adaptation in microbiome studies. DEBIAS-M learns bias-correction factors for each microbe in each batch that simultaneously minimize batch effects and maximize cross-study associations with phenotypes. Using benchmarks of HIV and colorectal cancer classification from gut microbiome data, and cervical neoplasia prediction from cervical microbiome data, we demonstrate that DEBIAS-M outperforms batch-correction methods commonly used in the field. Notably, we show that the inferred bias-correction factors are stable, interpretable, and strongly associated with specific experimental protocols. Overall, we show that DEBIAS-M allows for better modeling of microbiome data and identification of interpretable signals that are reproducible across studies.

bioinformatics↗

bioRxiv2022

Multilayer Networks of Plasmid Genetic Similarity Reveal Potential Pathways of Gene Transmission

Antimicrobial resistance (AMR) is a major threat to public health. Plasmids are principal vectors of antimicrobial resistance genes, greatly contributing to their spread and mobility across hosts. Nevertheless little is known about the dynamics of plasmid genetic exchange across animal hosts. The cow rumen ecosystem is an excellent model system because it hosts diverse plasmid communities which interact and exchange genes. Here, we use theory and methodology from network and disease ecology to investigate the potential of gene transmission between plasmids using a data-set of 21 plasmidomes from a single dairy cow population. We constructed a multilayer network based on pairwise genetic similarity between plasmids serving as a signature for past genetic exchange to identify potential routes and mechanisms of gene transmission within and between cows. The transmission network was dominated by links between cows. Modularity analysis unraveled a major cross-cow transmission pathway with additional small pathways. Plasmid functions influenced network structure: plasmids containing mobility genes were more connected; those with the same AMR genes formed their own modules. We find signatures of gene superspreading in which a few plasmids and cows are responsible for most gene exchange. An agent-based transmission model showed that a new gene invading the cow population is likely to reach all cows. Finally, we showed that link weights contain a non-random signature for the mechanisms of gene transmission allowing us to differentiate between dispersal and genetic exchange. These results provide insights into the mechanisms by which genes, including those providing AMR, spread across animal hosts.

ecology↗