Squeezing biology out of statistics: Gene set and pathway analysis in HT data
Berlin, 12-16 2018
Instructor: January Weiner (Staff scientist, Max Planck Institute for Infection Biology)
Course summary
High throughput (HT) techniques such as transcriptomics or metabolomics are of great significance in many areas of biology. However, the path from a boring list of differentially expressed genes to a biological understanding of the results is not straightforward.
This course offers computational techniques that go beyond a simple technical or statistical analysis. It covers techniques for the analysis of gene set enrichments, pathway analysis, gene ontologies, functional analysis of metabolomic profiling and making use of correlations and coexpression networks. A prominent part of the course will be devoted to data visualization and visual data exploration.
The students will gain the ability to independently process and analyse HT data sets, select the appropriate tools, functionally interpret the results as well as learn the paradigms of computational biology and statistics which will allow them to efficiently communicate with computational biologists.
As an incentive, each student will receive a set of gene expression profiles for a different organism, and during the course they will use these to generate species-specific gene expression modules and test their utility. If we are successful, we will attempt a joint publication.
Course prerequisites
In general, the course is aimed at biologists who would like to take their data analysis in their own hands. While an aptitude for computational work is necessary, the main goal of the course is the application of biological and statistical knowledge to HT sets with as little effort as necessary.
- basic computer skills (a rudimentary knowledge of programming principles in any language is recommended, but not mandatory)
- basic understanding of statistics
- basic understanding of molecular techniques for generating high throughput data
The students should be comfortable with using a computer and have at least a rudimentary understanding of computer programming. However, no specific skills are necessary; the students will learn basic R programming in this course.
Basic skills in statistics are necessary. The students should understand the concepts of statistical hypothesis testing and p-values. However, an in-depth introduction to these concepts will also be provided.
Target student skills
- understanding of computational problems associated with high-throughput data analysis
- statistical problems and solutions in functional analysis of HT data
- overview of commonly used functional analysis techniques (GSEA, gene ontologies, MSigDB, tmod, metabolic profiling)
- multivariate techniques and machine learning
- Communication skills in statistics and computational biology
After the course, the student should be able to prepare, analyse and functionally interpret a HT data set, including multivariate and machine learning techniques.
Course scheme
On each day, the course will consist of four parts:
- Lecture: theoretical introduction to the days focus
- Hands-on guide: guided practical session in R where students replicate the analysis performed by the teacher. While the lecture is general, here specific R techniques and R packages are introduced
- Guided self-study: students are given excercises and problems to solve and work on them individually under the guidance of the teacher
- Individual project work: each student will receive a transcriptomic (RNASeq or microarray) data set to analyse throughout the course
- Lecture: wrap-up and side notes; preparation for the following day
Course plan
Day 1: Introduction to statistical reasoning and R
- Lecture: "Statistics gone wrong: basics of statistical problems in HT applications"
- Hands-on guide: working with R: first steps
- Guided self-study: using R for data loading and basic statistical calculations
- Individual project work: loading data for the individual project
- Lecture: "So you have a list of thousand gene names: why do we do HT analyses?"
Day 2: Data preparation and a functional analysis primer
- Lecture: "Methods of pathway and functional analysis in gene set enrichment analyses"
- Hands-on guide: gene set enrichment techniques in R and other frameworks
- Guided self-study: comparing results of different gene set enrichment techniques
- Individual project work: biological interpretation of the results
- Lecture: "Common mistakes in functional analysis of HT data"
Day 3: Making your own modules
- Lecture: "Gene expression, co-expression and correlations"
- Hands-on guide: making your own modules
- Individual project work
Day 4: Multivariate approaches to functional analyses
- Lecture: "Introduction to multivariate approaches and ML techniques"
- Hands-on guide: Practical guide to multivariate and ML techniques in R
- Individual project work
- Lecture: "How to know when you are done?"
Day 5: Evaluation of individual project
- Individual work on project reports
- Project report evaluation
- Discussion round
- Lecture: "Course wrap-up: where to go from here?"
For more information about the course, please visit our website: https://www.physalia-courses.org/courses-workshops/course3/
Here is the full list of our courses and Workshops: https://www.physalia-courses.org/courses-workshops/
