September 2007-May 2008 Meetings
Friday, May 16, 2008 1-2pm Building 98B-1 Auditorium
Speaker: Dr. Aarti Shah, Lilly Executive Director
Topic: From India to Indianapolis; From a Number Cruncher to Executive Management
During this talk, Dr. Shaw will share her journey as a statistician. The focus will be on sharing her experiences - successes as well as failures. She will reflect on the following topics: transition from east to west, learning and adapting to different cultures, moving from technical to administrative leadership at Lilly and working in Corporate America as a woman. She looks forward to an engaging and interactive discussion. Please come to hear her top secrets which will make you successful in your career.
Thursday, March 13, 2008 9-10:15am Building 98 Auditorium
Speaker: Prof. Xiao-Li Meng, Chair, Department of Statistics, Harvard University
Topic: A Semi-Theoretician's Mid-Day Confession: The True Meaning of i.i.d. in (Applied) Statistics
Doing good (applied) statistics is inherently – and increasingly –difficult. The size of the data and the complexity of their structure are increasing, as are the depth and specificity of the investigation goals. Yet the available time for conducting the study is decreasing
due to intensive competition, especially for funding. Statistical consultants are therefore increasingly asked to perform magic, such as providing scientifically valid causal conclusions for a deeply stratified subgroup based on a handful of weighted samples with wildly varying weights. And by the way, the analysis must be done in one week
and the methods must be implementable by (and explainable to) analysts whose statistical experience might come largely from reading output from SAS/SPSS/Stata.
This is not a cynical observation, but rather a real challenge that we as statisticians must face in order for our profession to remain at the core of quantitative scientific investigation. Of course with every grand challenge comes a great opportunity! In this talk, I will report my own experiences both crying and smiling as a member of a team of statistical consultants for the National Latino and Asian American Study (NLAAS), a very recent survey of psychiatric epidemiology, which measured over 5000 variables and embedded experiments on different survey instruments. I will report on the success and failure of using Bayesian modeling and multiple imputation to deal with the respondents' untruthful self-reporting, as detected by the embedded experiments. If time permits, I will also show how the ambiguity in the Institute of Medicine's definition of "healthcare disparity" led to the concepts of conditional disparity and marginal disparity, which correspond to two extreme causal models. Yet the theory needed for the more realistic concept of joint disparity is still far from reach. The true meaning of i.i.d. will become clear only by the end of my talk; unless, of course, you have already deciphered it from this abstract….
Tuesday, October 23, 2007 Room 48-3-H3100 in Lilly Corporate Center
Speaker: Prof. Mary Ellen Bock, Department of Statistics, Purdue University
Topic: Current and Future Status of Statistics as a Profession
The increasing demand for statistical expertise has brought growing pains to the profession of statistics. The American Statistical Association, ASA, is the largest professional society for statisticians in the world. It includes members in industry, government and in academia and this talk is a discussion of the sources and effects of demand for statistical expertise on these three groups in the ASA. The talk includes some predictions for the future of the profession.
Friday, July 20, 2007
A Workshop by Geert Molenberghs consisting of a Pair of Short Courses: (1) The Statistical Evaluation of Surrogate Endpoints in Clinical Trials and (2) Incomplete Data in Longitudinal Studies.
Wednesday, May 16, 2007
Speaker: Prof. Jun Xie, Department of Statistics, Purdue University
Topic: Statistical methods for inferring gene regulatory modules and networks
A gene regulatory module is defined as a set of coexpressed genes that are regulated by a common set of transcription factors. In this talk, we focus on statistical approaches for inferring regulatory modules. We propose a series of statistical methods that combine information from multiple types of biological data, including genomic DNA sequences, genome-wide location analysis (ChIP-chip experiments), and mRNA gene expression microarray. More specifically, we developed a hidden Markov model to first predict combinations of transcription factor binding sites on DNA sequences. The predictions are refined by regression analysis on mRNA gene expression microarray data and/or ChIP-chip binding experiments. Our approach is validated on the well-studied yeast cell cycle gene regulation. The three data sources provide regulatory signals from different aspects. Therefore, the integrative analysis offers a better prediction on transcriptional regulatory code and infers potential regulatory networks.