11am - 12 noon
Thursday 29 April 2021
Integrating cell mechanics and cell mechanosensing, and uncertainty quantification for a time-series of count data and influence network
This event has passed
This seminar will take place online via Zoom.
This seminar comprises of two talks: Integrating cell mechanics and cell mechanosensing by Dr Carina Dunlop and uncertainty quantification for a time-series of count data and influence network by Dr Naratip Santitissadeekorn.
Integrating cell mechanics and cell mechanosensing
Dr Carina Dunlop
About the talk
Cells have been demonstrated to be extremely sensitive to the physical properties of their external environments, changing behaviours as diverse as proliferation, differentiation and migration in response.
The broad mechanisms by which this mechanosensing is achieved is broadly understood with cellular contractility shown to be an integral component. However, the mechanisms by which the physical coupling of a cell to its environment, cell contractility and signal transduction are integrated is less clear.
Here I present recent work on continuum elasticity models of cellular contractility and stiffness sensing. I show that this approach can predict cell shape dynamics, and provide an explanation for experimental observations on adhesion dynamics and mechanotransduction at the nuclear envelope.
Uncertainty quantification for a time-series of count data and influence network
Dr Naratip Santitissadeekorn
About the talk
A high-dimensional Hawkes process is widely used to model mutual excitation within a network where each node is described by a one-dimensional Hawkes process and the links represent the strength of mutual excitation.
For example, large-amplitude earthquakes could excite bursts of aftershocks within an earthquake network. Inference of a high-dimensional Hawkes process typically requires a time-stamp data. Hence, it is unavailable to a time-series of count data in general.
This talk will present my recent work that models the time-series of count data with discrete-time Hawkes processes as well as a sequential Monte Carlo (SMC) approach that recursively estimates the influence network from a times-series of count data.