Modeling and Inference of Transcriptional Gene Networks

ECS 289A, Spring 2004
CRN: 60541

Instructor: Vladimir Filkov
Time: TR, 9:00 am - 10:20 am
Place: 207 Olson
Units: 4

Course Description:

This is a seminar course on computational modeling and methods for inference of gene networks, using gene expression and other large-scale genetic data. The following discrete and continuous models will be discussed: graph models, Boolean networks, Bayesian networks, Linear models, and differential equations. The emphasis will be on data-driven, network inference methods, including combinatorial optimization, Bayesian reasoning, systems of linear equations, graph algorithms and others.

Goals:

  1. Introduction to modeling and model inference in large systems.
  2. Summarizing the state-of-the art in gene regulation modeling.
  3. Addressing problems presented by large-scale biological data.
  4. Emphasizing interdisciplinary research.

Prerequisites:

This course will cover topics from both biology and computer science, but the focus will be on algorithms and mathematical methods. The course will aim to be self sufficient, and will cover basic concepts, necessary for understanding of the material, from both disciplines. Graduate standing in Computer Science and/or Life Sciences, or permission of the instructor, will be required to register for the course. Graduate level familiarity with algorithms and data structures, as well as undergraduate preparation in numerical methods is recommended.

Expanded Course Description:

I. Gene Regulation and Technology

  1. DNA, transcription, translation
  2. High-throughput Genomic Technologies
  3. What is in the data and what is not?

II. Biological Properties of Gene Networks

  1. Biological Principles
  2. Cis-region logic and modularity
  3. Combinatorial interaction of transcription factors
  4. Gene Network Dynamics

III. Scientific Modeling Overview

  1. Physical and Combinatorial Modeling and Approximations
  2. Large-Scale Data Modeling

IV. Gene Network Models and Methods for Their Inference

  1. Combinatorial Models
    1. Graph (Static) Models and graph optimization methods
    2. Boolean Networks and Black-box, reverse-engineering algorithms
    3. Linear Models and Systems of linear equations
    4. Bayesian Networks, Bayesian reasoning and nonlinear optimization
  2. Continuous (Physical) Methods
    1. Differential Equations
    2. Canonical Models

Text:

Since there is no appropriate book in this area, a combination of lecture notes and recently published technical papers will be used.

Grading:

Students will be graded based on a term project (60%), two technical paper/software presentations (30%), and class participation (10%).

Project:

Students will use methods taught in class to follow the process of gene regulation inference from available data. Both theoretical and applied projects will be suggested. The projects will be done in groups consisting of a fair mix of life science and computer science students.

Presentations:

Students will present two technical papers or software used for gene regulation inference.