Modelling, clustering and aligning gene expression time series with probabilistic models

James Hensman (University of Sheffield)

Conference Room A, 2nd Floor, Schuster Building,

As methods for measuring gene expression become cheaper, it is becoming possible to monitor genome-wide expression in longitudinal studies, across development, disease progression or natural cycles. To deal with the complexities of these data we require models that are powerful, flexible and applicable to the resulting large datasets. I will show how hierarchical Gaussian process models can deal with biological and replicate variability of time series, and how such a model can be integrated into an efficient clustering scheme for genome sized data.  I'll present results from recent collaborations where such modelling has enhanced scientific study of the data.

Import this event to your Outlook calendar
▲ Up to the top