This document provides information about a graduate level bioinformatics course offered at the University of Washington. The course will introduce students to applying machine learning and computational methods to problems in biology and medicine. Over 12 weeks, topics will include biological foundations, machine learning algorithms, Bayesian models, neural networks, hidden Markov models, and more. Students will complete assignments, a group project, and exams. The goal is for students to understand bioinformatics problems and tools for solving them through various algorithms.

1. Bioinformatics TCSS 588A Winter 2007
TCSS 588 A Winter 2007
Bioinformatics
Instructor: Isabelle Bichindaritz, Ph.D.
Class: PNK 104 M/W 7:30 P.M. – 9:45 P.M.
CP 206M for assignments
Email: ibichind@u.washington.edu
Office: Cherry Parkes 216
Office hours: M/W 2:20 P.M. – 4:20 P.M.
always by email
by appointment
Class Web-site: http://courses.washington.edu/tcss588
DESCRIPTION
The bioinformatics course explains how to apply computer science to biology,
medicine, genomics, and proteomics. After a detailed presentation of a selection of
life sciences problems, the course examines and compares specific computational
methods and systems to solve them, focusing on machine learning, concept learning,
statistical learning, hidden Markov models, case based reasoning, neural networks,
evolutionary computing, knowledge-based systems and ontologies, stochastic
grammars and linguistics, grid computing, and semantic Web. Several applications
and projects (the Human Genome Project) are detailed, and tools to build new
applications are provided.
OBJECTIVES
Some of the objectives for this course include:
o Understand medical and biological concepts and set of problems.
o Understand scientific framework for bioinformatics in statistics, complexity, and
information theory.
o Understand machine learning algorithms and methods for biomedical informatics.
o Learn familiarity with specific machine learning methods and algorithms such as
statistical learning, concept learning, hidden Markov models, case based
reasoning, neural networks, knowledge-based systems and ontologies, genetic
algorithms, stochastic grammars and linguistics, grid computing, and semantic
Web.
o Compare different machine learning methods and learn how to select the best
suited to solve a particular problem in biomedical informatics.
o Program using available biomedical informatics tools (SPSS. Clementine, Weka,
Phylip, …).

2. Bioinformatics TCSS 588A Winter 2007
o Design and develop new computer systems for bioinformatics.
TOPICS
1. Biological and medical foundations.
2. Machine learning algorithms and applications to biology/life sciences.
3. Bayesian models.
4. Regression.
5. Multivariate methods.
6. Neural networks.
7. Hidden Markov Models.
8. Clustering.
9. Non parametric methods.
10. Decision trees.
11. Phylogenetic trees induction.
12. Combining several learners.
Detailed tentative schedule for each class, assignments, project, and schedules can be found at
the class home page at http://courses.washington.edu/tcss588 .
PREREQUISITE
Graduate students: TCSS343
Undergraduate students: Core completed.
TEXTBOOKS
Introduction to Machine Learning, Ethem ALPAYDIN, The MIT Press, 2004, ISBN
0-262-01211-1.
Medical Informatics: Knowledge Management and Data Mining in Biomedicine, Hsinnchun
Chen, Hsinchun Chen (Editor), Sherrilynne S. Fuller (Editor), Springer-Verlag, 2005, ISBN:
038724381X.
RECOMMENDED
o The Elements of Statistical Learning, Trevor Hastie, Robert Tibshirani, Jerome
Friedman, Springer Verlag, 2001, ISBN: 0-387-95284-5.
o Bioinformatics: The Machine Learning Approach (second edition), Pierre Baldi, Soren
Brunak, Sren Brunak, MIT Press, 2001, ISBN: 026202506X.
o An Introduction to Bioinformatics Algorithms, Neil C. Jones, Pavel A. Pevzner, the MIT
Press, 2004, ISBN: 0-262-10106-8.
o Inferring Phylogenies, Joseph Felsenstein, Sinauer Associates, 2004, ISBN:
0-87893-177-5.
o Data mining on multimedia data, Petra Perner, Springer Verlag, Series: Lecture Notes in
Computer Science, Vol. 2558, 2003, ISBN: 3-540-00317-7.
CLASS WORK AND EVALUATION

3. Bioinformatics TCSS 588A Winter 2007
There will be three individual assignments due, a project (group or individual), and two exams.
Assignments are due by midnight on the due date, and will be submitted electronically. The
deliverables will be either individual assignments (three assignments), or group project
deliverables. Assignments will be worked on in the lab. Assignments and the project may
involve programming, but design, writing, critical thinking, and analysis will be most
predominant. Project programming will be a team effort, if this option is chosen, with peer
evaluations due at the end of the quarter. Homework assignments and project deliverables are
posted on the class Web site. Incomplete assignments will be accepted. No late assignment will
be accepted.
PROJECT WORK
The project will be developed in teams of no more than 2 students. The teams will be formed
during the first two weeks of class. Although grades for the different components of the project
will be granted at the team-level, a peer evaluation will be performed at the end of the project,
and may alter, either positively, or negatively, a student’s final project grade.
PARTICIPATION
Participation will be evaluated by attendance, and active involvement in in-class discussions. In
order to participate efficiently in class, performing the pre-readings and critically thinking about
these is a prerequisite.
GRADING
Assignments: 25% (individual)
Project: 30% (individual or team)
Participation 15%
Midterm: 15%
Final: 15%
CODE OF CONDUCT
The assignments, and of course the quizzes, and exams need to be done individually. Copying
of another student's work or code, even if changes are subsequently made, is inappropriate,
and such work or code will not be accepted. The University has very clear guidelines for
academic misconduct, and they will be enforced in this class.
COURSE CHANGES
The schedule and procedures for this course are subject to change. Changes will be announced in
class and it is the student's responsibility to learn and adjust to changes.
IMPORTANT
If you would like to request academic accommodations due to a temporary or permanent
disability, please contact Lisa Tice, Manager of Disability Support Services (DSS) in the
Mattress Factory Bldg, Suite 206. An appointment can be made through the front desk of
Student Affairs (692-4400), through Student Services (692-4501), by phoning Lisa directly at
692-4493 (voice), 692-4413 (TTY), or by e-mail (ltice@u.washington.edu). Appropriate

4. Bioinformatics TCSS 588A Winter 2007
accommodations are arranged after you've conferred with the DSS Manager and presented the
required documentation of your disability to DSS.

5. Bioinformatics TCSS 588A Winter 2007
TENTATIVE SCHEDULE
Day Topic Assignment
Week Introduction to class
1 What is machine learning
Week Supervised learning Project proposal
2 Bayesian decision theory
Data mining in medical informatics
Week Classification / regression
3 Data mining in bioinformatics
Week Multivariate classification / Project bibliography
4 regression
Identification of biological
relationships
Week Dimensionality reduction Assignment #1
5 Clustering
Learning metabolic networks
Week Non parametric methods Evaluation plan
6 Decision trees
Learning gene pathways
MIDTERM
Week Neural networks Assignment #2
7 Multilayer perceptrons
The genomic data mine
Week Hidden Markov Models Draft of article
8 Exploratory genomic data analysis
Week Local models Assignment #3
9 Classification
Joint learning in genomics
Week Multiple learners Oral presentation
10 Reinforcement learning
Infectious disease informatics
Finals FINAL (M 3/12) Final article
week