"Research at Genentech is an opportunity to learn continually in the course of pursuing new treatments for patients while advancing biotechnology and science."
I trained originally in statistics at the University of California, Berkeley, with Professor Terry Speed. From there, I transitioned to a postdoctoral position in Human Genetics at the University of Pittsburgh with Professor Eleanor Feingold. After joining Genentech, I worked for several years in the Nonclinical Biostatistics group on projects ranging across manufacturing and quality control, nonclinical aspects of early clinical development, and basic research.
In 2014, I transferred into Bioinformatics & Computational Biology and currently lead a group of statisticians working closely with research scientists to learn from the torrents of data generated across modern biology. Examples of my group’s efforts include devising metrics and software tools for quality control and analysis of genetic sequences from clinical samples, developing a framework for modeling signals from circulating tumor DNA to monitor cancer progression, identifying and characterizing opportunities in functional genomics for improvements in both technology and methodology, and developing summary measures and software tools deployed in online applications to automate recurring complex statistical analyses.
Molecular medicine is an evolving and expanding body of knowledge and methods wherein each advance enables experiments that reveal new insights into human disease. I work closely with discovery and translational scientists, applying both classic statistical techniques and recent methodological advances to frame questions, design experiments, and craft & conduct statistical analyses that will extend our understanding so that the next round of investigation can come into focus.
Working in statistics within biotechnology is an opportunity to learn continually about diverse and developing areas. Some recent collaborations have included applying mixed models to proteomic assays to detect rare types of enzymes and advising bioinformatic scientists on choice and application of techniques from penalized regression and machine learning for identifying features of high-dimensional data sets relevant in cancer immunology. My current work is developing new statistical summaries and software employing regression splines to estimate changes in longitudinal data, with applications in translational oncology.