Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Bayesian-Guided Generation of Synthetic Microbiomes with Minimized Pathogenicity

Nisha Pillai, Bindu Nanduri, Michael J Rothrock, Zhiqian Chen, Mahalingam Ramkumar

TL;DR

The paper tackles the challenge of mitigating multidrug resistance (MDR) by generating synthetic microbiomes with minimized pathogenic potential. It integrates an autoencoder-based latent-space representation with Bayesian optimization to efficiently search microbiome designs, using Gaussian process models and acquisition functions (EI, UCB, PI, TS) to guide decoding of latent samples toward low-MDR predictions. Diversity-driven sampling via a determinantal point process and oversampling with SMOTE improve learning under limited data, while the latent-space decoder enables rapid generation and screening of synthetic microbiomes. The approach demonstrates reduced search iterations and identifies microbial taxa associated with lowered MDR, offering a pathway to bespoke microbiome designs with potential applications in poultry health and food safety.

Abstract

Synthetic microbiomes offer new possibilities for modulating microbiota, to address the barriers in multidtug resistance (MDR) research. We present a Bayesian optimization approach to enable efficient searching over the space of synthetic microbiome variants to identify candidates predictive of reduced MDR. Microbiome datasets were encoded into a low-dimensional latent space using autoencoders. Sampling from this space allowed generation of synthetic microbiome signatures. Bayesian optimization was then implemented to select variants for biological screening to maximize identification of designs with restricted MDR pathogens based on minimal samples. Four acquisition functions were evaluated: expected improvement, upper confidence bound, Thompson sampling, and probability of improvement. Based on each strategy, synthetic samples were prioritized according to their MDR detection. Expected improvement, upper confidence bound, and probability of improvement consistently produced synthetic microbiome candidates with significantly fewer searches than Thompson sampling. By combining deep latent space mapping and Bayesian learning for efficient guided screening, this study demonstrated the feasibility of creating bespoke synthetic microbiomes with customized MDR profiles.

Bayesian-Guided Generation of Synthetic Microbiomes with Minimized Pathogenicity

TL;DR

The paper tackles the challenge of mitigating multidrug resistance (MDR) by generating synthetic microbiomes with minimized pathogenic potential. It integrates an autoencoder-based latent-space representation with Bayesian optimization to efficiently search microbiome designs, using Gaussian process models and acquisition functions (EI, UCB, PI, TS) to guide decoding of latent samples toward low-MDR predictions. Diversity-driven sampling via a determinantal point process and oversampling with SMOTE improve learning under limited data, while the latent-space decoder enables rapid generation and screening of synthetic microbiomes. The approach demonstrates reduced search iterations and identifies microbial taxa associated with lowered MDR, offering a pathway to bespoke microbiome designs with potential applications in poultry health and food safety.

Abstract

Synthetic microbiomes offer new possibilities for modulating microbiota, to address the barriers in multidtug resistance (MDR) research. We present a Bayesian optimization approach to enable efficient searching over the space of synthetic microbiome variants to identify candidates predictive of reduced MDR. Microbiome datasets were encoded into a low-dimensional latent space using autoencoders. Sampling from this space allowed generation of synthetic microbiome signatures. Bayesian optimization was then implemented to select variants for biological screening to maximize identification of designs with restricted MDR pathogens based on minimal samples. Four acquisition functions were evaluated: expected improvement, upper confidence bound, Thompson sampling, and probability of improvement. Based on each strategy, synthetic samples were prioritized according to their MDR detection. Expected improvement, upper confidence bound, and probability of improvement consistently produced synthetic microbiome candidates with significantly fewer searches than Thompson sampling. By combining deep latent space mapping and Bayesian learning for efficient guided screening, this study demonstrated the feasibility of creating bespoke synthetic microbiomes with customized MDR profiles.
Paper Structure (17 sections, 4 figures, 3 tables)

This paper contains 17 sections, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Dataset
  3. Cultural Isolation for Antibiotic Sensitivity Testing
  4. Microbiome Analysis
  5. Approach
  6. Multi-Class Classification
  7. Diverse Point Selection
  8. Auto-Encoder
  9. Random Data Generation in Latent Space
  10. Bayesian Optimization
  11. Experiments And Results
  12. Multi-Class Classification
  13. Auto-Encoder
  14. Bayesian Optimization
  15. Synthetic Microbiome Selection
  16. ...and 2 more sections

Figures (4)

  • Figure 1: The learning phase of synthetic data generation architecture.
  • Figure 2: Generation of synthetic microbiomes using Bayesian optimization in few iterations.
  • Figure 3: Multi-class classification performance comparison at every iteration for Salmonella, Listeria, and Campylobacter MDR.
  • Figure 4: Top influencing microbiomes and their relative abundance values within a synthetic microbiome sample selected by our framework that predicts low MDR.