General Considerations

There is a lot of discussion about clinical research, but less about basic science (thus far). Vanderbilt is actually quite strong in translational research. One demonstration of this is the SPORE (NCI sponsored) research, which is extremely interdisciplinary and stresses translational research. Each SPORE project contains both basic science and clinical components (and each SPORE program has an epidemiology project). Vanderbilt has two on going SPORE programs in GI and breast cancer, and has had a lung SPORE as well. We are planning on submitting a brain and head and neck SPORE within the next couple of years. Biostatistics has been very strong in all of these SPOREs, and has always scored very high in the review process. Close by, there are many of us in the cancer center that have worked closely with bench scientists and have demonstrated expertise in statistical methods related to biologic techniques used in laboratory science.

A partial list:

Yu Shyr, Dean Billheimer, and Ming Li have been leaders in the application of statistical methods to proteomic measurement techniques (specifically MALDI-TOF mass spectrometry).

Yu Shyr and Bonnie LaFleur have published papers and been involved with statistical methods and software programs relating to microarray analysis.

Dean Billheimer and Bonnie LaFleur have done a lot of research related to methods of normalization used in many biologic techniques.

Dean Billheimer and Bonnie LaFleur have had experience in biomarker development. Dean is currently working on measurement aspects of biomarkers and assays and Bonnie has worked with drug development aspects (FDA) of biomarker and assay development.


More work in teaching important concepts used in biomarker discovery and validation as well as traditional concepts of measurement error and variation. (this is in addition to the need to teach more traditional – such as use in agricultural science – experimental design concepts).

Partial list of basic science programs that biostatistics can become more involved (this is related to question from Dan’s email: 4. What novel biostatistics and study design courses can we propose?  More efficient study designs.  Courses on the research program as a whole.)

IGP Program – Chun Li is already doing this with his course in human genetics.

Masters in Clinical Science Program (Ann Richmond and Ray Mernaugh) – a course for laboratory technicians who are getting a masters degree. This program was designed help people currently working at Vanderbilt promote their career without leaving their current positions to attend school full-time. It encourages retention in laboratory personnel, this is important since many labs spend a lot of time training people to do the specific techniques that may be unique to a particular lab.

These are the people who actually run (and sometimes design) experiments, so an understanding of measurement and basic experimental design would really benefit all basic science research at Vanderbilt.

MD/PhD Program – a course specific to students working toward a PhD in a basic science program (e.g., biochemistry, cell biology, cancer biology or bioengineering). These students take the first two years of medical school, and then take 2-3 years to obtain a PhD and then finish the last two years of medical school.

These students need help with basic research design, some complicated mathematical concepts, and statistical training. One thought I have had for these people would be a “clinic” or “brown bag” forum that would provide them with some direction for further study and specific training. It could be a mix of both lecture and discussion. The problem with “lecture only” is that their problems and questions vary greatly so flexibility would be essential. Some of these students have a strong mathematical background, though there are some that have some deficits in this area.

Medical Students – a course here might help increase physician-scientists receiving MSCI training (related to question 14. We had 14 new students in the MSCI program this year.  How can we increase the number of physician-scientists that receive this type of training?)

Having an elective course for 3rd/4th year medical students that would introduce clinical research and biostatistics methods would help them ready themselves for this career path. Just as with biostatistics education, the best time to “interest” future researchers is as early in their program as possible (e.g., high school/undergrad programs are the best place to interest the next generation of biostatisticians). Additionally, it would increase participation/visibility of biostatistics in the medical education program.

Verbage that can be added to current translational/basic science write-ups

Emerging measurement technologies provide unprecedented ability to measure individual components of complex systems. Many of these techniques share the characteristics of making highly multivariate measurements, but are limited in that only relative intensity is measured (often in arbitrary units). Nuisance variation affecting an entire observational unit (e.g., a gel, a microarray, a spectrum) is frequently present, and complicates ensuing data analysis. With these new assays, we require new quantitative methods; methods to account for inherent nuisance variation, and to aid interpretation of the assay in its biological/medical context.

Biostatisticians can provide quantitative expertise and assistance measurement techniques with respect to validity and reproducibility. We stress that quantitative methods development must reflect laboratory and biological knowledge. This requires close collaboration between laboratory scientists, clinicians and biostatisticians at all phases of assay development. A list, not exhaustive, of the relevant areas that can benefit from biostatistical proficiency are:

  • Baseline correction, multiplicative scaling, and transformations to improve distributional characteristics (e.g., a protein profile obtained from MALDI-TOF MS)
  • Units of measurement need to be clearly specified and analyzed appropriately (e.g., real time RT-PCR data)
  • Variation in repeated measurements from all biologic techniques needs to be assessed and analyzed appropriately. This includes both between sample variation and within sample variation.
  • Evlauation of normalization methods used in measurement techniques (e.g., microarray, blotting, and mass spectrometry data)
Topic revision: r2 - 26 Apr 2013, JohnBock

This site is powered by FoswikiCopyright © 2013-2022 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki? Send feedback