Computational Mathematics and Scientific Computing Seminar

Time and Location:

Speaker:

Link:

How do patterns in space and time emerge from molecular interactions in living systems? Answering this question is challenging even when the molecular participants in processes of interest are well-characterized because feedbacks are difficult to intuit and quantitative shifts in molecular features and populations can result in qualitative differences in patterns. Ever-increasing amounts of data now present the opportunity to evaluate models quantitatively, and simulation-based inference provides a principled approach. However, its use remains limited in cell biological contexts. I will discuss my group's recent efforts to use simulation-based inference---including recent advances that incorporate machine learning---to learn dynamical models of cell signaling in circadian and developmental contexts and show that simulation-based inference can reveal unanticipated mechanisms, in addition to quantitative insights.