Speaker 1: Ji Li

Title

Propensity Score and Bayesian Approaches to Leverage Real World Data

Abstract

In recent years, there is a growing interest in leveraging real-world data (RWD) in medical product development. Proper analysis of high-quality RWD can produce scientific evidence, make clinical trials more efficient and help medical decision making. To better borrow from external data, Bayesian paradigm provides a coherent approach: construct an informative prior (Power prior, Meta-analytic-predictive(MAP) prior, Elastic prior, etc.) and hierarchical modeling. Propensity score based approaches are also developed combining Bayesian inference. The simulations show that they have better performance than ordinary Bayesian approaches.

Speaker 2: Naghmeh Akhavan

Title

Migratory Border Cell Cluster Adaptations to Chemical Gradients in Fruit Fly Egg Chamber

Abstract

We mathematically model the migrating cluster in a 1-D spatial domain with variable cross-sectional area to represent the gaps on the cluster's route.  We use a 1-D reaction-diffusion PDE in a steady state with constant flux at one end to account for the chemoattractant distribution including at the gaps.  We model the movement of the cluster with a force balance question with forces at the front and back of the cluster driven by receptor dynamics responding to the local concentration of chemoattractant.  In this way, the cluster uses local information for cluster motion in the direction of higher chemoattractant concentration while slowing down through the gaps, as seen experimentally.