The Zhang laboratory reconstructs biological histories to enable better understanding of development, aging, and disease. To do this, Zhang lab develops novel computational and experimental methods operating across spatial and temporal domains at single-cell and single-molecule resolution. Many biological processes are dynamic, yet are typically accessible only through end-point samples or sparsely observed states. Zhang lab therefore designs methods to infer when and where molecular events occurred, including events that cannot be traced by conventional lineage-tracing or time-course approaches. A central theme of Zhang lab's work is the integration of observation with prior structure. Zhang lab views experimental data as embedded in a joint space defined by the measurements themselves and their latent biological priors. By jointly encoding observations and priors, Zhang lab approximates probability distributions that preserve biological signal and improve the recovery of latent biological states from high-dimensional data. Zhang lab is particularly interested in epigenomic inheritance. Cells within an individual multicellular organism share largely identical genomic sequences, yet acquire stable and diverse cell fates through epigenomic regulation. Imperfect propagation of epigenomic states during cell division contributes to aging, oncogenesis, and developmental disorders. Zhang lab combines single-molecule assays with statistical and machine-learning models to trace epigenome replication from parental to daughter molecules, to identify the molecular principles governing its fidelity.