CAREER: New Challenges in Statistical Genetics: Mendelian Randomization, Integrated Omics and General Methodology

点击翻译按钮获取中文摘要

With the rapid development of genetic technologies and the continuing collection of large-scale biobanks, scientists are provided with unprecedented opportunities to predict, prevent, and treat common diseases in a personalized and efficient way. In the meantime, analyzing such data presents many new challenges as 1) data come from multiple sources and can suffer from various biases and confounding; 2) scientific questions need an understanding of not only associations but also causal relationships among different risk factors and diseases. This project will address a range of statistical challenges in performing the integration of different omics data types to elucidate potential genetic changes that lead to disease development or relate to the discovery of treatment targets. The project will bridge statistics, machine learning, genetics, and medical research from an analytical perspective. In addition to helping young generations develop independent thinking, the educational activities will help them develop the ability to form objective opinions on social events and to analyze data to form an unbiased judgment on news stories. The PI will develop software and share research results on social media that can be useful to scientists and clinicians, doctors, and industrial professionals. This project also supports graduate students in the research. In the project, the PI plans to develop three aspects of research for modern statistical genetics. For Mendelian Randomization, which uses genetic mutations as natural experiments to understand risk factors for disease progression, the PI will focus specifically on adjusting for confounding and evaluating the temporal causal effects of clinical risk factors with new frameworks. For analyzing gene regulation with single-cell multi-omics data, the PI will work on a widespread regulation mechanism called alternative polyadenylation, building new statistical models to understand its functional roles with data from new technologies such as spatial transcriptomics and single-cell CRISPR screens. Furthermore, the PI will also investigate new statistical ideas motivated by recent methodological developments in genetics that can help to solve general problems in hypotheses testing and Bayesian inference.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.