Welcome to the phenotyping module of Bio5702, Current Approaches in Plant Research, a graduate course at Washington Unviersity in St. Louis. Over the next few weeks, we'll be exploring methods related to high-throughput phenotyping, including:
- Learning computational statistics with R
- Image analysis
- Morphometrics
- Multivariate statistics
You will be collecting your own data to analyze. The subject matter is: 75 different types of candy. In the next few weeks, you will measure 1000s of individual pieces of candy and analyze your dataset using multivariate statistics. The traits you will measure range from nutritional to morphometric to anything that you can thing of (including taste!).
This module was inspired by the video The Taxonomy of Candy by Emily Graslie. You can follow along on twitter for the course using the hashtag #CandyPhenotyping.
Check back here for updates on lecture topics and homework assignments.
The first lecture of this series will be an introduction to the statistical computing software environment R. In class, we will make sure everybody has downloaded R and go over the introduction together (script, PDF). The initial dataset we will use to learn R is the nutritional information of our 75 different candy types. At the end to the lecture, you should be able to:
- Read data into R
- Manipulate and format dataframes
- Create stunning visualziations using the ggplot2 package
- Perform descriptive statistics and initial data explroation
The beginning of this lecture we will go over the remainder of the ggplot2 tutorial. Your homework for this lecture is to create three graphs using ggplot2 that explore the nutritional candy dataset. We will go over your ggplot2 graphs in class next lecture.
For the remainder of this lecture, Chris gave a presentation introducing image analysis
The results from the homework assignment to make three ggplots using the candy_nutrition.txt dataset, the nutrition information for 75 different candy types, has been posted.
You can view the homework in your browser as html too.
Chris continued with the image analysis presentation and an introduction to image analysis in FIJI. The FIJI introduction is viewable as markdown or downloadable as a Word file.
Phenotyping went great today! Chris has posted all your photos to DropBox, but another place to find all the raw data is here. Homework for Friday is to cull duplicate/bad pictures and rename the files with the identification numbers before 10am on Friday. On Friday, you will use FIJI to make masks and gather data (feature extraction).
Chris spent more time lecturing about image analysis and students spent time in class learning about image stacks, batch processing, ROIs, masks, and isolating colorimetric and shape descriptor information for images in FIJI.
Introduction to Principal Component Analysis (PCA), hierarchical clustering, and other multivariate statistical methods. The lecture can be found here.
Class was spent going over the homework!!! Great job everyone!