Purdue Undergraduate Students Established a Statistical Framework for Studying Zebrafish Behavior
Purdue Undergraduate Students Established a Statistical Framework for Studying Zebrafish Behaviour
Zebrafish has become a popular research model for neurobiology and drug discovery. One reason for this popularity is that zebrafish embryos are small, which makes it easy for scientists to monitor the behaviour of many individual animals in parallel. In these behavioural studies, the zebrafish are often exposed to sudden environmental changes to trigger characteristic swimming behaviour. This behaviour is then measured over a long period of time under multiple trials. The resulting data structure is high dimensional and too complex to be analyzed by standard statistical tools. This has limited how much scientists can learn from the results. To address this problem, several undergraduate and graduate students in the Leung Lab collaborated with statisticians from the University of Georgia to establish a framework for analyzing high-dimensional behavioural data. These students first evaluated Hotelling's T-squared test, an advanced statistical tool for analyzing data obtained from many samples over a long period of time. In particular, the students used this test to analyze several typical situations that were commonly used in similar behavioural studies. Then, they determined the sample size needed for the statistical test to achieve desirable statistical power. Furthermore, they analyzed the intrinsic variations of the experimental design of the behavioural assay by multivariate analysis of variance. Together, their work has established a statistical framework for analyzing zebrafish behavioural data with similar complex structure. This framework will also guide the efficient design of zebrafish behavioural study to acquire critical information for advancing neurobiology and drug discovery.
This work has just been published in PLOS ONE.
The Purdue students involved in this study are:
- Robert Carmer. Robert is the co-first author of the paper. He was a statistics major who started working on this project under the support of The Howard Hughes Medical Institute Undergraduate Summer Research. He is now an associate in the Argus Information & Advisory Services, LLC.
- Gaonan Zhang. Gaonan was a biology major. He is now a PULSe graduate student at Purdue University.
- Prahatha Venkatraman. Prahatha is a graduate student in the Department of Biological Sciences. She is currently a Schlumberger Foundation Faculty for Future Fellow.
The Statisticians from University Georgia are:
- Yiwen Liu. Yiwen is the co-first author of the paper. She is a graduate student with Prof. Ping Ma.
- Ping Ma. Ping is a Professor in Statistics. He acquired his Ph.D. in statistics at Purdue University.
Liu Y, Carmer R, Zhang G, Venkatraman P, Brown SA, Pang CP, Zhang M, Ma P, Leung YF. Statistical analysis of zebrafish locomotor response. PLOS ONE. 2015; 10(10): e0139521.
Top row: Robert Carmer, Gaonan Zhang, and Prahatha Venkatraman;
Bottom row: Yiwen Liu, and Ping Ma.