Biometrics Section Announces JSM Program, Readies for Other Meetings

Edited by Songthip Ounpraseuth, Biometrics Section Publications Officer

The Biometrics Section will sponsor the following four short courses and six invited sessions during the 2012 Joint Statistical Meetings in San Diego, California:

Short Courses

• Statistical Analysis with Missing Data
  Instructors: Roderick Little and Trivellore Raghunathan, University of Michigan

  In this short course, we will discuss methods for the statistical analysis of data sets with missing values. Topics will include the following:

  1. Definition of missing data
  2. Assumptions about mechanisms, including missing at random
  3. Pros and cons of simple methods such as complete-case analysis and imputation
  4. Weighting methods
  5. Maximum likelihood and Bayesian inference with missing data
  6. Multiple imputation
  7. Computational techniques, including EM algorithm and extensions
  8. Gibbs sampler
  9. Software for handling missing data
  10. Missing data in common statistical applications, including regression, repeated-measures analysis, and clinical trials
  11. Selection and pattern-mixture models for nonrandom nonresponse

  Prerequisites: Knowledge of standard statistical models such as the multivariate normal, multiple linear regression, and contingency tables and matrix algebra, calculus, and basic maximum likelihood for common distributions

  Recommended text: Little, R. J., and D. B. Rubin. 2002. Statistical Analysis with Missing Data, 2nd edition.

• Smoothing Splines: Methods and Applications
  Instructor: Yuedong Wang, University of California at Santa Barbara

  This short course is about a particular class of modern nonparametric regression methods called spline smoothing for estimating functions of one and several variables. Special models such as polynomial, periodic, thin plate, partial, and tensor product smoothing splines for Gaussian data will be covered. The general form of smoothing spline models using reproducing kernel Hilbert space (RKHS) also will be discussed; however, no prior knowledge of these spaces is assumed. Data-based methods for estimating the optimal amount of smoothing such as cross validation, generalized cross validation, and generalized maximum likelihood will be explored. The short course will focus on methodology and application. We will provide a gentle introduction to the RKHS, keep theory at a minimum, and provide details about how the RKHS can be used to construct spline models. Much of the exposition is based on the analysis of real examples using R.

  Prerequisite: Two years of graduate-level statistics courses such as statistical inference, generalized linear models, data analysis, and regression.

  Recommended text: The short course will use partial material covered by Wang, Y. 2011. Smoothing Splines: Methods and Applications.

• Statistical Methods for Genome-Wide Association, Copy Number Variants, and Rare Variants Analysis
  Instructors: Hongzhe Li, University of Pennsylvania, and Wei Pan, University of Minnesota

  Our short course will introduce statistical issues and methods related to the analysis of genome-wide association data, copy number variation analysis, and analysis of rare variants and several important topics in human genetic research. Genome-wide association studies have been successful and have identified thousands of new genetic variants associated with common diseases. This course will introduce the rational, design, and standard analysis methods for GWAS and then move on to more advanced topics in genetic association studies that aim to identify the missing heritability. These topics include gene sets and pathway-based analysis of GWAS data, methods for copy number variants analysis, and methods for analysis of rare variants based on next-generation sequence data. Genome-wide association study of neuroblastoma will be used throughout this course to demonstrate these methods.

  Prerequisites: Participants are expected to have basic knowledge of statistical inferences such as hypothesis testing, regression modeling, and likelihood-based inferences. Also, they should have some familiarity with basic genetic terminology (e.g., alleles, genotypes, haplotypes, SNPs, linkage disequilibrium, etc.). Some knowledge of population genetics and molecular biology would be helpful, but is not required.

• Design and Analysis of Biomarker Studies for Risk Prediction
  Instructors: Tianxi Cai, Harvard University, and Yingye Zheng, Fred Hutchinson Cancer Research Center

  An accurate and individualized outcome prediction promises to dramatically change clinical decisionmaking in many branches of medicine such as early diagnosis of cancer and selecting patient-specific treatments. But translating the promise into reality is not easy. Clinical evaluations, while remaining an essential basis for risk assessment, may not be sufficient for complex diseases. Improved prediction may be achieved by combining information from multiple markers based on emerging new technology such as gene expression profiling, protein mass spectrometry, and proton emission tomography. Most marker tests are imperfect, and incorporate test results can have enormous consequences in both financial and human terms. Prior to incorporating a biomarker into standard clinical care, rigorous evaluation is required. Designing a rigorous study that efficiently uses available biologic specimens is critical. Compared to classical statistical methods for evaluating medical diagnostic test, there is relatively little literature devoted to statistical methods for marker development carried out in a prospective cohort study with censored failure time outcome. This short course will introduce recent statistical development for constructing and evaluating risk prediction model (markers) with censored data. While providing some mathematical details, we will emphasize the concepts, methods, and their real-world applications with the aim of both offering an overview of the rapidly developing area of risk prediction and biomarker evaluation and in-depth discussion about efficient design of biomarker and risk prediction studies.

Invited Sessions

• Recent Methodology Developed for the Design of Early-Phase Clinical Trials, organized by Thomas Braun
• Statistical Challenges and Innovative Solutions for Correlated Data, organized by Peiyong (Annie) Qu
• Statistical Methods for High-Dimensional Complex-Structured Object Data, organized by Veera Baladandayuthapani
• Biomarkers for Risk Prediction, Disease Detection, and Treatment Effect Estimation: Statistical Issues, organized by Layla Parast
• Shrinkage Estimation: Unifying Different Perspectives, organized by Bhramar Mukherjee
• New Methodological Advances in Network-Based Analysis of Omics Data, organized by Ali Shojaie

Mixer and Business Meeting

The section also will host a mixer and business meeting during JSM on July 30 from 5:30 p.m. to 7:00 p.m. The 2012 David P. Byar Young Investigator Award and section travel awards will be presented. The mixer is open to all JSM attendees and an excellent networking opportunity. Check the online program for updates about location.

ENAR 2013

It is time to think about invited sessions for ENAR 2013, which will be held March 10–13 in Orlando, Florida. Anyone who is interested in organizing an invited session or who has ideas for one should contact our 2013 program chair, Doug Schaubel, at deschau@umich.edu.

A typical session consists of three 30-minute talks followed by a 30-minute discussion or four 30-minute talks. June 11 is the deadline for proposals. It is best if you have a well-defined topic and commitments from participants by June 11. The more detailed the proposal, the better the chances it will be selected in this competitive process.

JSM 2013

It’s also time to start thinking about invited sessions for next year’s Joint Statistical Meetings, which will be held August 3–8 in Montréal, Québec, Canada. Anyone interested in organizing an invited session or who has ideas for one should contact our 2013 program chair, Wei Sun, at wsun@bios.unc.edu.

A typical invited session consists of three 30-minute talks followed by a 10-minute invited discussion and 10 minutes of floor discussion; however, other formats are possible. The 2012 program is a good source for examples.

Remember, the most mature ideas will have an advantage, so it’s best to have your ideas in final form by the middle of June. The Biometrics Section will have at least four invited sessions, but if we generate enough good ideas, we will be able to compete for additional slots.

Also, submit ideas for short courses to our 2013&endash;2014 continuing education chair, Donglin Zeng, at dzeng@email.unc.edu.