BS Animal Science National Taiwan University; MA Animal Science Louisiana State University; Ph.D Animal Breeding Texas A & M University; Biostatistician Hoffmann La Roche Pharmaceutical Inc.

How would an animal breeding major become a biostatistician? As an Animal Science major I went into the field of Quantitative Genetics. I found myself taking more and more statistics courses. It fascinates me to find out how much modern agriculture development is attributable to experimental statistics and how many basic statistical theories were first developed by animal and crop geneticists. Basically, I love math and life sciences so biostatistics is the perfect marriage.

After graduate school, I designed experiments to develop veterinarian medicine for both farm animals and pets for seven years. Now I work in the same capacity in human medicines. Why would discovering a new medicine have anything to do with statistics? Well, in order to know whether the drug works, we have to test it first on animals and then on patients who volunteer to be tested.

Often the results are not black or white. One question is how many patients need to be tested before we know a drug is effective and safe. How do we know these results are not purely due to chance? These are all statistical questions. I help the clinicians find the best way to conduct an experiment to answer the most important questions to them. I calculate the number of subjects needed in each trial in order to make a conclusion with a given certainty on the effect of a new drug.

What is needed to be an effective member of a team depends mostly on common sense and interpersonal skills. Math gives you a solid background for analytical thinking. No schooling prepares you for the rest. You have to make it up as you go. Therefore, you have to love what you're doing.

When I worked in the USDA Forest Experimental Station as a junior statistician many years ago, I did not have any statistical theory courses. For several weeks, I was plotting some samples of tree diameter distribution. I discovered there were some patterns in the ratio of mean over standard deviation. Later my colleague and I published a paper on a quick way to estimate the parameters for the Weibull distribution. And, much later, I found out that I had just reinvented the Moment Method which was known many decades before. But I still think it's great because I had so much fun doing it.

Your career should be something you do and have so much fun doing that sometimes you feel a little guilty about taking the money.