Exploratory analysis and downstream analysis

Keywords

Statistical-model, Exploratory-analysis

Authors

Charlotte Soneson (@charlotte), charlottesoneson@gmail.com

Type

• Lecture

Description

This lecture gives an overview of exploratory analysis (clustering) and supervised analysis (prediction/classification), as well as visualization methods (heatmaps/PCA) and gene set analysis. It also shows how to transform count data to make it more suitable to apply the traditional methods developed (e.g.) for microarray data.

Aims

The aim of the lecture is to introduce the audience to exploratory analysis, to show how to easily obtain an initial visual overview of the data, and to perform functional enrichment analysis of differential expression results. The aim is also to point to particular characteristics of RNA-seq data, and how to transform the data to be more amenable to classical analysis methods.

Prerequisites

• Know what a count matrix represents
• Some background on differential expression analysis (at least conceptually)

Target audience

beginner, biologist

Learning objectives

• Using visualization to identify confounding effects and taking necessary action
• Building a simple classifier and apply it to data
• Choosing an appropriate strategy for gene set analysis
• Performing gene set analysis
• Interpreting the output

Materials

• Lecture slides

Data

• The Bottomly data set (downloaded from ReCount) is used to create some of the slides.

Timing

Approximately half a day of lecture

Content stability

The content is relatively stable.

Technical requirements

• Not applicable

Literature references

• Not applicable

Comments

• I did not check if the use of all figures is allowed or properly acknowledged.
• A license needs to be added