By Chris Dwan
April 24, 2005 | Curriculum change is making its way through biology departments at universities across the country. Much as physics was transformed by computing nearly 50 years ago, biology educators are now integrating computer programming, cluster computing, and even the fundamentals of databases into their undergraduate coursework.
This change is well motivated: In 2003, the National Research Council published “Bio 2010 — Transforming Undergraduate Education for Future Research Biologists.” The goal of its committee was to examine undergraduate education in the life sciences. In the introduction to their report, they point out:
In contrast to biological research, undergraduate biology education has changed relatively little during the past two decades. The ways in which most future research biologists are educated are geared to the biology of the past, rather than to the biology of the present or future.
The report makes several provocative suggestions, including a statement that medical school admissions requirements are anchoring biology curricula in the past and preventing needed change. One less controversial suggestion is that courses be incrementally modified to include modules that address new or cross-disciplinary material. In fact, this report served as a catalyst for such module development and the growing trend of interdisciplinary training.
The BioTeam thinks that effectively extending this training to undergraduates and eventually to high-schoolers will require innovative uses of the interface to provide just the right level of power and complexity.
High-performance computing systems are complex, and the Unix command line is unwieldy and difficult for the uninitiated. This is doubly true in the context of an academic module lasting no more than a week at the undergraduate level. Anyone who has taught or attended a one-day workshop on the Unix command line can testify that a single day is barely enough to cover vocabulary, much less empower exploration.
Educators need to provide a structured yet limited introduction to informatics that inspires interest without intimidating the students. Professors need to be sufficiently comfortable with the technologies they are using to answer questions that arise. This combination of requirements, in our opinion, removes the Unix command line as an option for short introductory modules.
Hamilton College, in Clinton, N.Y., has integrated a bioinformatics cluster into its Bio 111 course. Instructors developed highly customized interfaces to a few popular sequence analysis tools, including BLAST, CLUSTALW, and PHYLIP. They did this using a Web interface to their cluster in which information on inputs, outputs, and interface display is encoded in XML. A variety of interfaces are then automatically prepared based on that XML. This allows the authors of the XML to reveal as much or as little complexity as they desire.
In the Hamilton example, students start with a highly structured tutorial interface. Most parameters are pre-set to a specific default. Instead of using the documentation spaces of the interface to explain which choices are available, it provides a high level of detail about the particular choice that was already made for the student, complete with references to the course text. As students progress, they advance to a simplified interface with actual choices for the parameters. The full interface is available for those who want to make use of it.
Web interfaces, such as those developed at Hamilton, are the most flexible and accessible way to provide less sophisticated users with access to centralized computing resources such as clusters. In many cases, once the user realizes that there is a huge payoff in terms of computational power for using a cluster, he or she will take the time required to learn the mysteries of the command line.
As we see resources such as cluster computers moving out of the supercomputing “glass house” and into more labs, high schools, and offices, automation, ease of use, and interface design will become more critically important than they already are.
Chris Dwan is a senior consultant with The BioTeam. E-mail: cdwan@bioteam.net.