Coordinating Organizer
Breanna Shi
Shi00231@umn.edu

This workshop will focus on the application of Bayesian Networks (BNs) to count data originating from next-generation sequencing experiments. The following specific topics will be addressed:

  • Introduction to Bayesian statistics and probabilistic models
  • Comparative discussion of BNs packages in R
  • Generating and testing hypotheses using BNs
  • Applications of BNs to metabolomic and microbiome data


Target Audience

Undergraduates,graduate students and postdocs with an interest in learning about Bayesian statistics and applying Bayesian Networks to microbiome data.


Organizer(s)

Breanna Shi


Christoper Dean


Jesse Elder



Session Plan (Four Hours)

Introduction(30 mins)

  • Introduce the mathematical basis of Bayesian Statistics
  • Importance and utility of networks

General Applications(1 hour)

  • Bayesian Networks - How do these ideas combine into a relevant application?
  • Comparison of techniques for applying Bayesian Networks (R, Python, Matlab)
  • Discussion: The challenge of choosing priors
  • Second Introduction: Microbiome & Antimicrobial Resistance (AMR)
  • Brief overview of empirical Bayesian Statistics
  • Different approaches to BNs (for metagenomic data, specifically)

Interactivity(2.5 hour)

  • Walkthrough of Meta AMR pipeline
  • Use Microbiome count data make a network using Hill-Climbing algorithm in R
  • Use output of Hill-Climbing algorithm to make a Gephi file
  • Create a network in Gephi software
  • Use Bayesian theorem to find the strength of significant connections

Learning Objectives

  • What is Bayesian Statistics and how can it be used in real-world scenarios?
  • How BNs are applied to ‘omics data?
  • Step-by-step how to apply BNs to AMR count data
  • Utilize Bayesian networks to understand microbiome and AMR relationships