Research News
The Department of Biostatistics has continued its tradition of strong methodological and collaborative research in 2022. Collectively, Biostatistics faculty have published more than 425 papers in high-profile journals in biostatistics, public health and medicine. We received five new methodological awards as PI and MPIs, bringing the total number of on-going biostatistics research grants to twenty six. Biostatistics faculty also collaborated on 136 grants in public health and medicine as co-investigators. To accelerate and further motivate the generation of federally funded cutting-edge methodology grants, we have instituted a new internal grant mechanism for grant writing. In its inaugural year, the program has funded two pilot grants to groups led by Dr. Jeff Goldsmith and Dr. Yifei Sun. Several issues related to the department’s research mission were discussed at our department’s day-long retreat on September, 2022, giving rise to a list of areas that need attention. A number of working groups are currently deliberating on prioritizing the many recommendations with aim towards tangible action.
Thanks to the great efforts of the recruitment committee led by Drs. Zhezhen Jin and Jeff Goldsmith, one of the most memorable achievements of the year 2022 has been the successful recruitment of four amazingly talented faculty members, Drs. Wenpin Hou, Molei Liu, Zhonghua Liu, and Xiao Wu. Their cutting-edge research energizes the department and opens new front lines in our research and collaborations. All except for Dr. Xiao Wu, who will join us in January 2023, have already joined us over the summer.
In the past year, Dr. Wenpin Hou developed several innovative analytical tools for single-cell spatial transcriptomics, from computation to visualization. Single-cell spatial transcriptomics is a new ground-breaking molecular profiling technique that allows scientists to measure all the gene activities in a tissue at the single-cell resolution and map out the spatial locations of those activities. Its promise for future biology and human health can only be fully realized with proper analytical tools. Dr. Hou’s well-rounded analytical framework delivers one solution.
Dr. Zhonghua Liu and his co-authors recently developed a new robust Mendelian randomization method called MR MiSTERI by leveraging phenotypic variance to identify a causal effect correctly. Mendelian randomization uses genetic profiles as instrumental variables. It is a widely used approach to overcome a major limitation of observational studies: unmeasured confounding. Dr. Liu’s MR MiSTERI is powerful, flexible, and robust to the violations of traditional core assumptions. As a result, it can broadly empower the use of modern biobank databases for scientific discoveries and evaluation of interventions. Key targets include the UK biobank, the All of US research program, and many institutional-level biobanks.
With the advances in cloud computing, data collection on all fronts and storage in different geographic locations can now be accessible to researchers. Dr. Liu has developed several practical tools which enable the analysis of such connected data in a federated and distributed fashion, and have attracted considerable attention. One of his recent contributions is a distilled conditional randomization test (dCRT), which allows the use of a wide range of state-of-the-art machine learning algorithms to test for conditional independence. This methodological breakthrough has great potential for scientific discovery and facilitates causal inference with complex machine-learning methods on large data sets.
Dr. Xiao Wu dedicates his methodological research to addressing urgent health research needs in the context of the environment and climate change. One such example is a new generalized propensity score approach which he developed last year to facilitate causal inference for continuous exposures. This new method fills a gap in the existing causal inference literature and can promote awareness of causal inference in future science and policy-relevant research in environmental and climate studies, where interventions and exposures are naturally continuous.
Recruiting new talent is one of the department's main tasks in building strong, modern, and connected biostatistics and health data science research. In the coming years, we will continue to lead methodological advances in systematic learning from large-scale and connected data resources for scientific discoveries and precision invention learning. We will build more bridges to connect with informatics, computer science, and engineering, and to promote the core values of statistical rigor, principles, and reasoning in advancing health data science. We will also continue to strengthen the connections among our efforts in research, education, and training, to accelerate knowledge transfer to empower future biostatisticians and the healthcare workforce generally. And finally, and crucially, we will enhance our connections with underrepresented and underprivileged communities, using modern data science approaches to mitigate health disparities and promote health for all. With these features in mind, we look forward to another exciting and productive year of research in 2023.