Statistical Analysis of Mouse Gut Microbiota

1
Statistical Analysis of Mouse Gut Microbiota

• Alex Tran
• Computer Systems Lab 2009-2010

2
Abstract

• The mouse gut microbiotal community is the
  population of bacteria, which inhabit the
  digestive track of a given population of mice.
  The ability to understand the microbiotal
  community has implications, which extend beyond
  mice and into factors of human obesity and the
  dietary needs of the human body. Computational
  genomics is an emerging field, which blends both
  biology and computer science, to analyze genes.
  This study seeks to utilize several methods
  emerging within computational genomics to analyze
  the gut microbiome of mice to observe the effects
  of varying diets on the gut microbiotal
  community, as well as create a universal and user
  friendly tool for researchers to utilize when
  studying any gut microbiome. The applications of
  this can extend past studying the gut microbiome,
  but can also be applied to any taxonomic group
  being counted for analysis, however several parts
  of the application are exclusively beneficial to
  the analysis of gut microbiota specifically those
  found in mice.

3
Introduction

• Genomics
• Gut Microbiota
• Mouse Digestive Tract

http//gchelpdesk.ualberta.ca/news/02jun05/images/
circos-001.png
An example of a Taxanomic tree.
4
Background

• OB/OB mice
• Human Microbiome Project

5
Washington University Gordon Lab

• Working with Gut Microbiota in Obese Mice
• Specifically observing Fecal matter
• Larger Sample Size
• Mice last longer
• Currently pay 1,000 dollars for each batch of
  tests
• Comes back in an unreadable format
• Biologists can't program
• All grad students are forced to learn Perl

6
Objectives

• User Friendliness
• Efficiency
• Universal Compatibility
• Hugenholtz
• RDP
• Ability to load any Taxanomic tree

7
Development

• Python
• Tkinter
• I/O intensive
• cPickle

8
1st 2nd and 3rd Quarter Progress

• 3rd Quarter
• Shell for detailed analysis implemented
• Customized Tkinter File Browser added
• Archiving with recently used files and trees
  ready
• New combobox ability implemented
• Ability to have parallel programs communicate is
  ready for implementation

• 1st Quarter
• Very Basic Gui shell in place
• Ability to read 2 taxonomic trees and strictly
  formatted files
• No exception catching ability
• No user selected taxonomies
• No clear data analysis possible
• 2nd Quarter
• Driver and application style implementation
• Shell for multiple file loading and comparison
• User friendly error catch system
• Reads standard format files (with error
  detection)
• Any Taxonomic Tree

9
Final Results

• Ability to specify taxonomy used
• Basic statistical comparisons implemented
• Multiple file loading implemented
• Use in Gordon Lab
• Crossfile simultaneous OTU selection
• Initial loadscreen for tree selection and hub
  loading

10
Going Forward

• From the ground up approach
• More in depth statistical analysis
• Ability to view more than just a handful of files
  at a time
• More user friendly archiving