** Co-first authors
Publications:
Dynamic Subgroup Identification in Covariate-adjusted Response-adaptive. Li, Y., Wang, J. and Wei, W. (2024). Neural Information Processing Systems (NeurIPS).
Proposed a dynamic subgroup identification covariate-adjusted response-adaptive randomization design strategy that dynamically identifies and merges the best-performing subgroups based on collected data to maximize experimental efficiency.
Demonstrated through theoretical investigations that the proposed design has a higher probability of correctly identifying the best set of subgroups compared to conventional designs.
Proved the statistical validity of the estimator for the best subgroup treatment effect with its asymptotic normality and semiparametric efficiency.
Validated the proposed design using synthetic data from a clinical trial on cirrhosis.
Preprints:
Uncertainty quantification in epigenetic clocks via conformalized quantile regression. Li, Y., Goodrich, J. M., Peterson, K. E., Song, P. X. K. and Luo, L. (2025). Genetic Epidemiology.
Developed a pipeline integrating high-dimensional quantile regression and conformal prediction to train epigenetic clocks, effectively revealing population heterogeneity and constructing adaptive prediction intervals.
Demonstrated that quantile regression-based prediction intervals are narrower and more statistically efficient than those from conventional mean regression-based clocks with DNA methylation datasets from children.
Revealed cellular evolutionary heterogeneity in age patterns in different developmental stages during individual childhoods and adolescent cohort.
Provided valuable insights for future applications of epigenetic interventions for age-related diseases.