MEBI 550: Winter 2010 course information

Check here for homework assignments, course notes, and other hopefully useful stuff.

Tentative lecture topic schedule

The topics we will cover are tentatively scheduled, though we may slip or add or substitute other topics. The schedule is available as a PDF.

Course texts

• Kalet, Principles of Biomedical Informatics (required), referred to below and in other course documents as "PBI". As the author, I receive a small royalty on each book. I have established an endowed fund at the UW, the Ira and Terry Kalet Fund for Biomedical Informatics Trainees, using this royalty income. Future royalties will also go into this fund, not retained by me personally.
• Graham, ANSI Common Lisp (required), referred to as "Graham".

Computing resources

The Informatics computing lab (I-Lab) of the Biomedical and Health Informatics Graduate Program, located in T-277, Health Sciences Building, provides a Linux server and desktop systems with a Common Lisp programming environment. To obtain an account and learn about access, after registering for the course, contact Gary Csorgo (gcsorgo@u.washington.edu), the Informatics Lab Manager.

Lecture slides and notes

This list will be expanded as the course progresses.

• Course intro, with a little about biomedical data representation.
• Biomedical data, tagged data, databases and XML.
• Search methods, strategies and applications
• Logic intro, propositional logic, and implementation of a theorem prover
• First Order Logic, introducing variables and quantifiers
• Expert systems, medical examples, guideline systems
• Alternatives to First Order Logic, including frames
• Description Logics
• Probability and reasoning under uncertainty
• Information theory
• Machine learning

Reading assignments

Reading assignments are on the course schedule (see above), at the end of the topic list for each date. They are specified by chapter and section, so for example, PBI 1.2 means Chapter 1, Section 2, in "Principles of Biomedical Informatics". Reading assignments should be completed by the indicated class date, in order to be prepared for class. Some of the readings are available through the UW Electronic Course Reserves (Eres) system, at the MEBI 550 page.

Homework

Homework assignments are listed here. Unless otherwise noted, homework is due by the beginning of class on the due date.

• Homework 1, due Tuesday, January 26 (revised from earlier). You will need code from PBI, Chapter 1, available at the PBI web page or the Common Lisp code page on the course web site. You will also need the data file, brca1.txt, and some files from raatkirani, the I-lab Linux server.
• Homework 2, due Tuesday, February 2.
• Homework 3, due Tuesday, February 16. No programming this time, just proofs.
• Homework 4, due Tuesday, March 2.
• Homework 5, due Thursday, March 11.

Solutions to Homework

Solutions will be emailed

Final Project

This year, instead of a final exam, students will be expected to do projects, and submit written reports.

Here are titles for some projects of students in previous years:

• Building a Knowledge-Based System to Support Patient-Centered Care for Diabetes
• Computational Analysis of the Cell Regulatory Cycle using the BioBike Biological Knowledge Base and Lisp Interface
• Analyze the genome of some interesting species to see whether the frequency of occurrence of an amino acid in the proteins correlates with the number of redundant codons for the amino acids, i.e., do the amino acids with 4 to 6 codons occur more frequently than the ones for which there are only one or two codons?

Here are some ideas for projects that have not been done before:

• Apply the state-machine simulation code in Chapter 8 of PBI to some other gene regulatory or protein interaction network, to see what properties of the network you can predict.
• Build an ontology of some biomedical or health area you are interested in, using any of the existing ontology building tools, e.g., Protege, Ontolingua, LOOM, etc.
• Apply an existing knowledge resource to answer some complex biomedical question, e.g. use the Gene Ontology to classify microarray data as in Jeff Shrager's "intelligent microarray analysis" tutorial.
• Get Wanda Pratt's DynaCat program working and apply it to some additional test cases (see instructor for details).