Syllabus: Applied Statistics for High-Throughput Biology
Levi Waldron, PhD
Assistant Professor of Biostatistics
City University of New York School Graduate of Public Health and Health Policy
New York, NY, U.S.A.
Times and Places
Classes will take place on March 6, 8, and 13:
|6 Marc 2018||09:00 - 11:30||Gamma [Borgo Roma - Ca' Vignal 2]|
|8 Marc 2018||09:00 - 11:30||G [Borgo Roma - Ca' Vignal 2]|
|13 Marc 2018||09:00 - 11:30||Gamma [Borgo Roma - Ca' Vignal 2]|
Please come to the first class with the following installed:
- Bioconductor www.bioconductor.org/install
- R Studio: https://www.rstudio.com/products/rstudio/download3/
This course will provide biologists and bioinformaticians with practical statistical and data analysis skills to perform rigorous analysis of high-throughput biological data. The course assumes some familiarity with genomics and with R programming, but does not assume prior statistical training. It covers the statistical concepts necessary to design experiments and analyze high-dimensional data generated by genomic technologies, including: exploratory data analysis, linear modeling, analysis of categorical variables, principal components analysis, and batch effects.
- Github resources at http://waldronlab.github.io/github/
- Resources for learning R at http://waldronlab.github.io/learnr/
- Other resources at http://waldronlab.github.io/
Each day will include a hands-on lab session, that students should attempt and hand in something before the following class by committing to this Github repository. You are encouraged to work together on lab exercises, but should hand in your own individual work.
A project will be handed out before the final class, that will involve analysing a genomics dataset. Each student will analyse their own dataset and prepare an individual report using R Markdown for reproducible analysis and reporting. Reports will be assessed for quality of analysis and clarity of presentation.
Session detail by day
All course materials will be available from https://github.com/waldronlab/AppStatBio/.
- random variables
- hypothesis testing for one or two samples (t-test, Wilcoxon test, etc)
- hypothesis testing for categorical variables (Fisher's Test, Chi-square test)
- data manipulation using dplyr
- linear modeling
- linear and generalized linear modeling
- model matrix and model formulae
- multiple testing
- unsupervised analysis
- graphics for exploratory data analysis
- distance in high dimensions
- principal components analysis and multidimensional scaling
- unsupervised clustering
- batch effects