BME 225 –
Protein Function in Biology and Bioinformatics
Please direct any enquiries to Dietlind Gerloff (gerloff@soe.ucsc.edu; 9x4833; PSB320)
Course description
Course reviewing functional roles of proteins and computational methods used to predict functional aspects of proteins. The primary focus will be on molecular function (catalysis of biochemical reactions; physical interactions with DNA, RNA, or other proteins; ion channels etc.) and structure-function relationships. Wider-reaching notions of function (pathways, interaction networks) will be considered peripherally, as the context in which molecular function occurs. The course is taught as lectures, (computational) lab work, and discussions of topical publications. Students not satisfying the prerequisite requirement may seek instructor's permission to enroll (with additional reading assignments if necessary).
We welcome graduate students (and senior undergrads) from all departments and schools - if you do not fulfill the prerequisites but are interested in the subject, please get in touch with the instructor. Much of the course is taught in discussion style, the more varied the backgrounds the more interesting for everyone!
Students interested in this course may also
be interested in BME281G (2 credits):
PROTEIN FUNCTION JOURNAL CLUB /
Gerloff group meeting
Course times and location
MON & WED 5:00-6:45pm in PSB305 (Physical Sciences Building)
As the course progresses, topics and copies of the slides will be posted here.
An overview schedule with assignment due dates can be found here.
A list of function prediction servers and programs for the projects/presentations can be found here.
Wed Dec 10, 11:30am: Project Presentations + a little bit about the Evolutionary Trace Methods (papers distributed last week)
(Provisional) Course Outline and Structure
The course will generally involve one taught session (lecture) and one applied/practical session (computational lab exercises; paper discussions; student presentations) per week.
The course is loosely structured in a biological part (overviewing the diversity of natural protein function on behalf of examples, and the wetlab methods used for functional characterization) and a bioinformatics part (reviewing current strategies and methods to predict function).
The topics tenatively scheduled for each week are:
By comparison to protein structure the term “protein function” is far less well defined. Researchers will use the term vaguely and differently depending on their background and interests. What schemes are there to sub-classify the wide variety of functions carried out by proteins in their natural environment - what are their strengths and weaknesses?
Week 2 + 3: Enzymes
The best understood type of molecular function is catalysis of a biochemical reaction. We also know the 3D-structures of many enzymes. Can we understand their mechanisms (better) by looking at these structures – are there commonalities that may be useful for predicting enzyme function?
Interactions between proteins and DNA and/or RNA are crucial to all transcriptional regulation. Best understood and characterized are specific interactions, e.g. recognition of a promoter sequence by a transcription factor. Additionally there are a variety of more “generic” DNA/RNA-binding events that are of great importance, e.g. for discarding unwanted RNA molecules etc. Do 3D-structures provide enough information to “break the recognition code”? How much constraint is imposed onto evolutionary sequence variation by this type of function?
Due to an (even) greater variety in shape and physicochemical properties in protein partners, the diversity in protein-protein interactions is almost ungraspable (for us bioinformaticians). Issues to consider and discuss here are: transient versus permanent interactions, conformational changes (including those triggered by post-translational modifications) , specific versus unspecific interactions. Where may lie clues for prediction methods basing on sequence and/or structure?
There are many other types of protein function that may be considered variations or combinations of the types discussed previously, or not. Our knowledge of molecular protein function is generally limited to “well-behaved” proteins, i.e. those adopting well-defined 3D-structures, typically at a minimum energy state. What other proteins are there and are their functions not also molecular in nature? What clues toward protein function can be gained by proteomic (experimental high-throughput) methods?
Where these have not been mentioned during the previous weeks, a selection of existing computational strategies to tackle protein function prediction will be reviewed and discussed. Why is prediction of function so difficult? This part of the course will include student presentations and discussions guided by the instructor.
Evaluation
Students will be evaluated based on their performance on
1. Two homework assignments (15% each)
2. An individual project involving literature and WWW/computational research, evaluated based on a summary report/paper (max 10 pages; research to be carried out before Week 8) (40%) - NOTE: senior grads (i.e. advanced grads) get a reprieve and don't have to write a report.
3. Discussion of one topical publication (journal club-style) (20%)
4. Participation in class (10%)
Reading Sources
No one text covers all of the information needed for this class. Much reading will be scientific publications in contemporary journals, and other sources available online.
However, students taking this course are encouraged to accompany their studies with selected chapters from the following texts:
• Protein Structure and Function
Gregory Petsko & Dagmar Ringe
New Science Press, 2003
Gaurav Pandey, Vipin Kumar & Michael Steinbach
To be published by Wiley Press (expected Fall 2008).
Alternatively, the technical report forming the basis of the book can be used:
http://www.cs.umn.edu/tech_reports_upload/tr2006/06-028.pdf