Personal tools
Home Rätsch Lab Talks & Lectures Machine Learning in Bioinformatics

Machine Learning in Bioinformatics

Course on "Machine Learning in Bioinformatics" at the Machine Learning Summer School.

Machine Learning in Bioinformatics

Abstract:

I will start by giving a general introduction into Bioinformatics, including basic biology, typical data types (sequences, structures, expression data and networks) and established analysis tasks. In the second part, I will discuss the problem of predictive sequence analysis with Support Vector Machines (SVMs). I will introduce a series of kernels suitable for different analysis tasks. Furthermore I will discuss the basic data structures needed for large scale learning and how to combine kernels for heterogeneous data. In the third part, I will focus on Hidden Markov models and discriminative alternatives like Conditional Random Fields and Hidden Markov SVMs suitable for segmentation tasks frequently appearing in Bioinformatics. In the last part I will present three applications in greater detail: A large margin alignment algorithm, computational gene finding and the identification of polymorphisms from resequencing arrays.

Overview:

  1. Introduction to Bioinformatics (45 min)
    • Basic Biology and Central Dogma
    • Typical Data Types
    • Common Analysis Tasks
  2. Sequence Analysis with SVMs (105 min)
    • String Kernels
    • Large Scale Data Structures
    • Heterogeneous Data
  3. Structured Output Learning (30 min)
    • Hidden Markov Models
    • Dynamic Programming
    • Discriminative Approaches
  4. Some Applications (90 min)
    • Spliced Alignments
    • Gene Finding
    • Analysis of Resequencing Arrays
Document Actions