Lesson 5: T-distributed Stochastic Neighbor Embedding

Table of Contents

Lecture 5

5.0 Lesson learning objectives

By the end of this lesson, we should understand what is t-distributed stochastic neighbor embedding and how to apply it to our spatial transcriptomic datasets.

Hands-on component 5

Our in-class hands-on component will analyzing our spatial transcriptomics datasets to create a data visualization using tSNE. We will also use gridExtra to create a multipanel plot.

Class Lesson Notes 5

Prof. Fan’s whiteboard notes from class: Please consult Slack and OneDrive

Prof. Fan’s code from class: code-02-03-2023.R (click to download)

Homework Assignment 5

Make a new data visualization of your spatial transcriptomics dataset with a minimum of 2 panels using the dimensionality reduction approaches we have been learning about in class. In particular, use data visualization to explore one of the following questions:

  1. Should I normalize and/or transform the gene expression data (e.g. log and/or scale) prior to dimensionality reduction?
  2. Should I perform non-linear dimensionality reduction on genes or PCs?
  3. If I perform non-linear dimensionality reduction on PCs, how many PCs should I use?
  4. What genes (or other cell features such as area or total genes detected) are driving my reduced dimensional components?

Write a description describing your data visualization using vocabulary terms from Lesson 1.

Use the same process from HW1 for what is expected and how to submit your homework to the folder hw3/. Additional details: genomic-data-visualization-HW_3.pptx (click to download)

Make a pull request to submit your HW3 (due Monday Midnight).

HW2 due today at Midnight.

Additional resources