What happens when a statistician is handed over a sample size of one?
Poulami Barman, a statistician at Mayo Clinic, USA faced a similar challenge while working on a prostrate cancer trial. In this guest blog piece, Poulami writes about her journey from Vellore Institute of Technology (VIT) in India to the premier US lab, how she handled the transition and the uniqueness of working on a sample size of one.
Till high school, I was this nerd who loved both math and biology. After 12th grade, I realised I was not “nerd enough” to pursue a career in biology and decided to go the math route. I wanted the best of both worlds and chose biotechnology for a major in undergrad — it slowly became clear to me that I was wired more towards the math equations and didn’t enjoy the biology side that much. With little career advice, I thought a complete change in major was not possible, and hence got into a Masters in Bioinformatics in the US.
During the first year of grad school, all my nightmares came alive in the form of coding and the biology knowledge needed for course work. The culture shock did not make it easy and I was almost dropping out of school but stayed on when my father insisted. I got an internship at a reputed hospital — that further exposed my weak coding skills. The turning point, however, came when I got an assistantship with Dr. Yolande Tra. She recognised my skill in statistics and encouraged me to take more courses in this field.
Under Dr. Tra’s mentorship, I aced in statistical analysis class which helped me land my first full-time job as a Clinical Data Analyst at Johnson & Johnson. My excitement did not last long, as I soon realised that the job involved a lot of coding and not much statistics. I did not like the monotony of the job and decided to go back to school, but this time to get a second masters in statistics. I moved to Texas A&M University – where I enjoyed working 14-16 hours a day and realised what passion actually meant. I got a job at Mayo Clinic after Masters, and there was no looking back. I never settled for a career until I had exactly what I wanted.
Statistics in medical science
With the growing complexity of biological data and disease biology, clinicians are leaning more towards personalised care therapies. The field of oncology is no stranger to this. 4 years ago, I got involved in one such study to recruit stage IV prostate cancer patients that are castrate resistant and to find specific markers that cause drug resistance. This was to be done by sequencing their genomic data before and after Abiraterone acetate/prednisone (AA/P) treatment. The presence or absence of markers could then decide the treatment regimen for the patients.
Data was collected with trials on mice injected with tumour cells from patients. Various treatment combinations were tested on the mice models. It was like having parallel patient avatars built from mice. A similar trial was conducted at Mayo Clinic on breast cancer patients. I was lucky to be involved in both these projects at some point. This is one of the best perks of being able to work in this specific role.
The pet peeve of all statisticians is a small sample size. And I was asked to provide inferences with just one sample! The analysis needed to focus more on gene-set within patient comparisons or one-vs-many control comparisons. The ideas was to learn the association of time-to-treatment change. We found that in metastases, Wnt/β-catenin pathway activation is associated with primary AA/P resistance and increased CCP with acquired drug resistance.
It has been a unique study in many ways, and to name a few; it provided a mutational genomic landscape for metastatic prostate cancer; mice models of metastatic prostate cancer (which has never been done before); and a stepping stone to single sample (n=1) analysis. The findings from the PROstate Cancer Medically Optimized Genome Enhanced ThErapy (PROMOTE) study was recently reported in PLOS One in 2015.
Bending the rules
Being involved in interesting and challenging projects like this one, and being paid to do what I love is a dream come true. The first couple of years at Mayo, I suffered from the Imposter Syndrome. I was working closely with researchers, and highly qualified people, and only 2 masters without a doctorate seemed petty. Don’t let the misconception of requiring a doctorate to pursue a career in research deter you. As long as one has the persistence and the penchant for it, research need not be sour grapes.
From an unsure novice in the US to someone who works with the best researchers in her field, I have grown a lot. And it was not a straight line; in fact, it was the worst and the best roller coaster ride I have ever taken. Every time I see someone eating street food at 2 a m in a Bollywood movie, and every Dushera, I miss home terribly. But, I think I have made a home for myself here — a small group of friends, and if I had to call any place home, other than Kolkata, it would be my Texas family. I met my husband and now have a sweet little family here in Rochester MN.