ABSTAT18

Course Description

This course is targeted for Biostatistical techniques often employed in analytical tools for high throughput data and multivariate data. Participants can expect to attend a thorough set of lectures that will reveal the conceptual frameworks that are needed to understand the methods. Extensive hands-on practice will be the main vehicle for providing the skills and user independence. To keep things in context, the course is exclusively based on biological examples. We will be using custom-built R scripts and packages that are available from the CRAN and/or Bioconductor repositories. Care has been taken not to use any proprietary data or software, so that the hands-on experience can carry on after the course, providing maximum user independence. We will be using custom-built R scripts and packages that are available from the CRAN and Bioconductor repositories.

Target Audience

Everybody using Bioinformatics methods is implicitly using statistical methods. Moreover, proper judgment of the results often calls for a deeper level of understanding than what is required to solve scholarly exercises. We will look into particular areas such Simulation, Bayesian Inference, Hidden Markov Chains and Multivariate Data Analysis methods with the attitude, eyes and brains of an experienced statistician that wants to understand how the methods work and systematic way.

Detailed Program

Days	Lectures	Exercises & Script examples
Day 1	Quick Review of Statistical Concepts Monte Carlo and Bootstrap methods Guides R Cheat sheet R Reference Card Probability Distributions in R	Probability Review (exercise) Random variables Parameters Discrete Distributions Continuous Distributions Simulation (exercise) Monte Carlo simulation Parametic/Non-parametric Boostrap estimation Boostrap Confidence Intervals Bootstrap Hypothesis Tests for one sample/two independent samples
Day 2	Bayesian Inference Gibbs Sampling Expectation-Maximization algorithm	Baseyian Inference Scipt example Gibbs Sampling (exercise) Scipt example Expectation-Maximization algorithm (exercise) Scipt example
Day 3	Principal Component Analysis Multiple Testing Long Version Short Version	Principal Component Analysis (exercise) Scipt example Multiple testing (exercise) Script exercises Script examples
Day 4		Project Expectation-Maximization (exercise) Project Principal Component and Bootstrap2 (exercise)