Data Analysis in the Biological Sciences: A 10-week hands-on course to equip students with tools for quantitative cell biology

The big questions in cell biology increasingly require quantitative approaches to get answers. A typical graduate student spends nearly as much time analyzing her data as she does acquiring them. Most graduate students, and indeed even senior researchers, learn analysis techniques as they need them. This often results in choosing the most familiar or commonly used analysis tool, as opposed to the optimal one. To meet the growing needs of students as they begin their careers in quantitative biology, I developed a hands-on course covering basic data analysis techniques useful in cell and developmental biology. The goal is build a repertoire of functional skills with a basic understanding of the underlying mathematics. The target audience is senior undergraduates and beginning graduate students of biology and bioengineering. Each of the ten weeks of the course consists of a one hour lecture covering the theoretical background for the topic of the week, and a three hour "lab" session in which the TAs and I work with the students to analyze real biological data sets, including published results from literature and unpublished results from labs on campus. Each week's homework also features analysis of real data. Topics include regression, parameter estimation, outlier detection and correction, error estimation, denoising, hypothesis testing, image processing and quantification, and data display and presentation. (Absent is analysis of sequence data, for which Caltech offers a separate bioinformatics course.) We approach statistics from a Bayesian perspective and use Python as the instructional programming language. I will present some of the examples of data sets we analyzed in the course, discuss the effectiveness of the course structure, and provide student feedback