Distinguished Scientist Marian Fisher to Retire from BMI

Distinguished Scientist Marian Fisher to Retire from BMI

It is with deep regret and the most heartfelt of best wishes that we announce that Marian Fisher will be resigning and retiring from her post in the Department at the end of February, 2016.  She is currently Research Professor and Distinguished Scientist as well as able Director of the Statistical Data Analysis Center within the Department, a post she has held for 24 years. We are truly grateful and indebted to her for her many years of dedicated, thoughtful and rigorous service to SDAC, the Department, and the field of clinical trials.

A research program within the BMI Department, the SDAC is a leader in promoting statistical practice, applications, and research in the design and analysis of clinical trials, especially in the preparation of Data Monitoring Committee reports necessary to evaluate the accumulating evidence for safety and efficacy over the course of a trial. Its primary collaborators are in the pharmaceutical industry. Dr. Fisher is to be credited with building and maintaining excellent connections with several leading drug development firms during her tenure. The SDAC’s work has lead not only to a strong reputation for the Department, but also serves as a locus of local expertise in clinical trials available to the SMPH faculty and an impetus for methodological research by SDAC investigators, and provides regular opportunities for BMI and Statistics graduate students to gain invaluable training experience in the design, conduct, and analysis of clinical trials.  Associate Scientist Kevin Buhr will assume the post of Director of SDAC, with Senior Scientist Tom Cook serving as Associate Director, starting on March 1.

Dr. Marian Fisher completed a BS in Mathematics at the Case Institute of Technology, and a MS and PhD in Mathematical Statistics at, respectively the University of Iowa and George Washington University.  She served as Mathematical Statistician and Health Statistician at the NHLBI, was Assistant Professor at the University of Maryland School of Medicine, and was Senior Statistician at the Maryland Medical Research Institute, all before joining BMI in 1990.  She has served the world’s clinical trials participants through many many appointments to data monitoring boards, and has served as PI on over 15 major industry contracts here at UW. Her publications include results from major trials and important methodological contributions as well.

Please note that we will seek an appropriate time and place to properly celebrate Marian’s career and, in the meantime, join me in thanking her for her service and in wishing her the very best in this next phase of her life!

Colin Dewey is 2016-2018 Vilas Associate

BMI Associate Professor Colin Dewey has been chosen as a 2016-2018 Vilas Associate!

The Vilas Associates Competition of the UW-Madison recognizes new and on-going research of the highest quality and significance. Recipients are chosen competitively by the divisional Research Committees on the basis of a detailed proposal.

Colin was selected for this award by the Division of Biological Sciences for work entitled "Accurate de novo transcriptome assembly by optimization with a biologically and statistically-motivated objective function.” As has been the hallmark of Colin’s work, this project reflects a deep understanding of the biology of the problem, together with a precise specification of an appropriate statistical model. The expected result is a transcriptome assembler that achieves high accuracy by explicitly optimizing a statistically and biologically-principled objective function.

Prof. Eneida Mendonca on AMIA Board of Directors

BMI Associate Professor Eneida Mendonca has been elected to a two-year term (January 1, 2016 – December 31, 2017) as Secretary on the Board of Directors of the American Medical Informatics Society. From its website, "AMIA is the professional home of leading informaticians: clinicians, scientists, researchers, educators, students, and other informatics professionals who rely on data to connect people, information, and technology. … It is the center of action for more than 5,000 health care professionals, informatics researchers, and thought-leaders in biomedicine, health care and science.”

Congratulations to Eneida and thank you for your service in this premier organization!

Symposium on Big Privacy: Policy Meets Data Science

The Center for Predictive Computational Phenotyping (CPCP) and the NIH Center of Excellence for Big Data Computing is pleased to announce this symposium on the privacy of big data.

Symposium sponsored by
The Center for Predictive Computational Phenotyping (CPCP)
NIH Center of Excellence for Big Data Computing

October 15th, 2015
1:00 PM - 5:00 PM
Discovery Building on the UW-Madison Campus
DeLuca Forum

With the advent of high-throughput methods in biomedical research, the drive for precision medicine, and the advances in computational methods that foster “big data science,” many commentators have expressed concern about how to promote biomedical science while respecting people’s privacy. Biomedical research data may be subject to different privacy laws and regulations depending on the type of institution holding or using the data, the type of data, who funds the research, the state in which the research is conducted, and other factors. Biomedical researchers are generally required to protect patient and research participant privacy, while at the same time researchers are encouraged or explicitly required to share data with the scientific community. In some cases privacy protections can impede science, but in some cases data sharing can expose research participants or patients to informational risk. This half-day symposium will examine legal, policy, and technical issues at the intersection of data privacy and data science. 

Registration is mandatory for both attending the symposium and to receive webcast information.
Please register by Wednesday, October 7th.
Please use this link to register:

Asst. Prof. Yajuan Si to Work on Unified Theory for Survey Sampling

Congratulations to Asst. Prof. Yajuan Si, jointly appointed in Biostatistics and Medical Informatics and in Population Health Sciences, on the receipt of her first NSF grant!

The project aims to develop a unified framework for survey weighting through novel modifications of multilevel regression and poststratification (MrP) to incorporate design-based information into modeling.

Real-life survey data often are unrepresentative due to selection bias and non-response. Existing methods for adjusting for known differences between the sample and population from which the sample is drawn have some advantages but also practical limitations. Classical weights are subject to large variability and can result in unstable estimators, while regression approaches present computational and modeling challenges. The new framework developed by these investigators will allow adjustment for selection bias and non-response as well as improvements in design-respecting inference.

Using this approach, survey analysts will be able to properly account for unignorable design issues in the regression framework, and practitioners who conduct surveys in government, academic, commercial, and non-profit sectors will be able to construct statistically efficient survey weights in a routine manner. This new framework may be applicable to problems resulting from the newly emerging explosion of "big data", such as integration of surveys from multiple sources, analysis of streaming data, and respondent-driven sampling.

The project will develop software that can be accessed by the general research community.

New Technique Can Help Researchers Explore the Rhythm of Genes

Oscope uncovers dynamic signals of technical origin in scRNA-seq data sets

MADISON, Wis. — Genetic cycles, from circadian rhythms to the cycles involved in reproduction, are everywhere. Now, a recent University of Wisconsin-Madison study is putting a new tool in the arsenal of researchers when it comes to the ebb and flow of gene expression. The new statistical approach, named Oscope, lets researchers identify and characterize the rhythm of genes across the entire genome using single-cell RNA sequencing.

Common genetic cycles include the circadian rhythm that regulates our behaviors over the 24-hour period and those that govern the division and multiplication of cells. Cycles are evident across almost all living systems, and errors in the process often lead to a variety of diseases. To study these cycles using traditional technologies, researchers must synchronize a whole population of cells so they are at the same state. Unfortunately, such synchronization isn’t possible for many cell types and conditions.

However, new RNA sequencing technology allows scientists to probe the genome wide expression of a single cell. When the cell is harvested for sequencing, it is destroyed in the process so it can’t be sequenced again to uncover an oscillating gene pattern. But with that single-cell information, researchers were able to ‘reorder’ unsynchronized cells and uncover a pattern of expression.

Oscillating gene expression can be thought of like a sine wave, moving up and down on a graph as the gene is expressed, repressed, and expressed again over time. Oscope uses a dataset from unsynchronized, single-cell RNA sequencing.

Christina Kendziorski, a professor of biostatistics and medical informatics at the UW School of Medicine and Public Health, led the team that developed Oscope to address the drawbacks of existing methods used to explore gene expression. The project melded the statistical strengths of the Kendziorski lab with the cell biology expertise from the lab of James Thomson, pioneer stem cell biologist at the UW and researcher at the Morgridge Institute for Research in Madison.

“This project is a great example of the collaborative efforts that are required to address the most challenging problems facing scientists today,” said Kendziorski.

Techniques to study oscillating gene expression in individual cells have existed for a long time. They include fluorescent tagging, where a certain gene is ‘tagged’ with a light-emitting molecule and gene expression is measured in proportion to the light emitted by the tag. If the tag is emitting light, the gene is active. However, this technique is labor-intensive and can only examine a handful of genes at a time.

When genetic rhythms are of interest across the entire genome, researchers often use traditional or bulk RNA sequencing, where a researcher examines the average gene expression over a large population of cells at various times. But to look at different times, the technique requires multiple populations of cells because the population is destroyed when expression is measured. The other issue here is the populations must be synchronized—starting and continuing at the same state of expression—which doesn’t always hold over time.

“After looking at average gene expression for over a decade, the ability to see genome-wide gene expression in individual cells is particularly exciting. Unfortunately, we only have a snapshot, and that’s the tricky part. We want to study oscillatory gene expression, but we don’t have time course data,” said Kendziorski, “so we developed a statistical method that would allow us to look at oscillatory genes and reconstruct one cycle of their oscillation that doesn’t involve time course experiments or synchronization.”

While the technique offers a new way to study gene oscillations, it does have limitations. The method can only reconstruct one cycle of the oscillation instead of its full oscillatory path. Another limitation is that with time course, one can see much more detailed dynamic for many genes.

Yet, for those working with oscillatory genes but uninterested in studying their rhythms, Oscope may be helpful in weeding out unnecessary data, letting a researcher adjust for the genetic variance that comes with oscillation, thereby improving the power to see the signal they’re most interested in.

This study was published in Nature Methods on August 24, 2015. The work was supported by US National Institutes of Health grants GM102756, 4UH3TR000506 and 5U01HL099773, the Charlotte Geyer Foundation and the Morgridge Institute for Research.

Professor Kim to present at Univeristat Autònoma de Barcelona

KyungMann Kim, Ph.D., Professor of Biostatistics and Statistics, has been
invited for a two-week (October 4-18) research visit to the Centre de Recerca
Matemàtica (CRM) located at the Universitat Autònoma de Barcelona as a Simons
Visiting Researcher. This will be Dr. Kim's second visit to the Universitat
Autònoma de Barcelona where he gave a two and a half day short course entitled
"Data and Safety Monitoring and Interim Analysis in Clinical Trials: Sequential
Methods" in June 2001.

The Simons Visiting Program is funded by the generous support of the
Mathematics and Physical Sciences branch of the Simons Foundation, the
Universitat Autònoma de Barcelona, and the Universitat Politècnica de
Catalunya, and its aim is to enhance the CRM Intensive Research Programme in
Statistical Advances for Complex Data.

Prof. Richard Chappell is 2015 AbbVie Statistics Visiting Scholar

Biostatistics and Medical Informatics Professor Rick Chappell has been named the 2015 AbbVie Statistics Visiting Scholar. In this role, Dr. Chappell will spend considerable time over the coming year in-house collaborating with AbbVie statisticians and other biopharmaceutical scientists on problems related to statistical design and analytical methodology needed to fully exploit information collected from AbbVie clinical trials. Whereas the primary focus will be on the design and analysis of AbbVie trials, from concept through analysis and reporting, Rick may also conduct seminars, engage in training and mentoring activities, and/or participate in the design of protocols or charter templates for novel types of clinical trials or other research. AbbVie ( is a global, research-based biopharmaceutical company situated in North Chicago.

Cook and Buhr Present Novel Course on Valid Clinical Trials

Senior Scientist Tom Cook and Associate Scientist Kevin Buhr, in collaboration with University of Utah Associate Professor Charlie Casper presented an evening workshop titled “Conducting Valid Trials: The Critical (and Misunderstood) Roles of Randomization and Complete Follow-Up” at the Society for Clinical Trials 36th Annual Meeting in Arlington, Virginia.

Randomized, double-blind clinical trials are widely considered the gold standard for establishing causal relationships between interventions and outcomes of interest. However, the importance of randomization and the intention-to-treat (ITT) principle in the design, conduct, and analysis of such trials — while widely recognized — is often not well understood. In this first-time SCT workshop, Cook and colleagues used easily understood models and examples, accessible to a non-mathematical audience, to demonstrate why randomization and ITT analysis lead directly to valid assessments of whether a new treatment “works”.

Despite some initial skepticism on the part of the Society's Education Committee, the workshop was very well received and sported the second highest attendance at the meeting. The presenters are planning to extend it to a longer morning workshop for next year's meeting.

The Society for Clinical Trials is an international professional organization dedicated to the development and dissemination of knowledge about the best practice in design, conduct, analysis and reporting of clinical trials with membership from government, academia, industry and non-profit organizations.

Associate Professor Menggang Yu Approved for over $1M Research Funding Award by the Patient-Centered Outcomes Research Institute

"The left panel represents a convex approximation for the 0-1 loss. In particular, the approximation is a surrogate logistic loss where the parameter "a" controls the shape of the surrogate. The right panel represents a non-convex approximation for the 0-1 loss where the parameter "a" controls the proximity of the surrogate."

Congratulations are in order to Associate Professor Menggang Yu for his imminent funding award from the Patient-Centered Outcomes Research Institute (PCORI) for his project entitled "Matching Complex Patients to Treatments: Innovative Statistical Scoring Methods for Treatment Selection". The project, which will run for 3 years, aims to develop and disseminate innovative statistical methods to rank possible treatments for patients according to their likelihood of achieving desirable patient outcomes. The ranking will be based on patient characteristics such as demographics and socio-economic variables, inpatient/outpatient diagnoses, comorbidity, pharmacy claims, health system and clinic information.

The award is one of four or five PCORI awards received recently by investigators at the University of Wisconsin School of Medicine and Public Health, and highlights the Department’s commitment to health services research.

The study is one of 46 proposals that PCORI approved for funding on Tuesday, April 21, to advance the field of comparative clinical effectiveness research (CER) and provide patients, healthcare providers, and other clinical decision makers with information that will help them make better-informed choices.
The award has been approved pending completion of a business and programmatic review by PCORI staff and issuance of a formal award contract to the University.

“This project was selected for PCORI funding not only for its scientific merit and commitment to engaging patients and other stakeholders, but also for its potential to fill an important gap in our health knowledge and give people information to help them weigh the effectiveness of their care options,” said PCORI Executive Director Joe Selby, MD, MPH. “We look forward to following the study’s progress and working with the University of Wisconsin School of Medicine and Public Health to share the results.”

Professor Yu’s study and the other projects approved for PCORI funding were selected through a highly competitive review process in which patients, clinicians, and other stakeholders joined clinical scientists to evaluate the proposals. Applications were assessed for scientific merit, how well they will engage patients and other stakeholders, and their methodological rigor among other criteria.

PCORI is an independent, non-profit organization authorized by Congress in 2010 to fund comparative clinical effectiveness research that will provide patients, their caregivers, and clinicians with the evidence-based information needed to make better-informed health and healthcare decisions. PCORI is committed to seeking input from a broad range of stakeholders to guide its work.

PCORI has approved $854 million to support 399 research studies and initiatives since it began funding research in 2012. For more information about PCORI funding, visit