18.417 Lecture 14

Announcements

Regulatory Networks: Example

Regulatory Networks: Approaches

Regulatory Networks: Goal

Expression Clustering

The problem

Expression Profiles

Clustering expression levels

1. Expression Value Normalization

2. Feature extraction

3. Clustering Algorithms

3a. Hierarchical clustering

3b. K-means clustering

3c. Cluster Representation

3d. Cluster distance metrics

PPT Slide

Other clustering methods

Evaluating clustering output

Visualizing clustering output

Rearranging tree branches

What have we learned?

Location Analysis

The question

Footprint experiments

Chromatin IP (ImmunoPrecipitation)

Location Analysis

What have we learned?

Sequence Conservation Isame gene across species

The idea

PPT Slide

Applying Bayes’ Law

Motif divergence and evolutionary time

PPT Slide

PPT Slide

Motif degeneracy and Random Occurrence

Signal to Noise in Multiple Species

PPT Slide

Sequence conservation IIacross co-regulated genes

The idea

The Algorithms

Suffix Trees (review)

PPT Slide

Gibbs sampling ideas

EM: Expectation Maximization

Implementation

From Clusters to Motifs

Sequence Conservation

Bayes Networks

Modeling the dependencies

Bayesian network topology

PPT Slide

Evaluating Alternative Hypotheses

PPT Slide

Model Comparison

Scoring all possible models

Summary

Regulatory Networks: The methods