Gaussian Processes and Gene Regulation
Gaussian Processes (Rasmussen and Williams, 2005) can
provide a convenient framework for the analysis of gene regulation as
shown in two recent papers (Lawrence et al., 2007; Gao et al., 2008).
In these two papers, the expression levels of a set of genes were
known and they were governed by a common regulatory factor (TF).
The unobserved concentration of TF was described by a Gaussian
Process. This approach has several advantages and should be
explored to the full. Relative to the papers above three useful
generalisation immediately springs to mind: Firstly, Multiple TFs.
Secondly, Network Models of GR in a single organism, that will allow
the TFs to be products of the genes whose expression levels are
observed. Thirdly, Network Models and Model Organisms will in the
simplest case occur, if experiments in for instance human and mouse
have been performed on homologous sets of molecules. A model of the
evolution of GR would be needed.
Gaussian Processes (Rasmussen and Williams, 2005) can provide a
Printer Friendly Version 