Classical models of intracellular signalling describe how small changes in
a cell's external environment can bring about major changes in cellular act
ivity. Recent findings from experimental biology indicate that many intrace
llular signalling systems show a high level of spatial organisation, This p
ermits the modification, by protein kinase or protein phosphatase action, o
f specific subsets of intracellular proteins - an attribute that is not add
ressed in classical signalling models. Here we use ideas and concepts from
computer science to describe the information processing nature of intracell
ular signalling pathways and the impact of spatial heterogeneity of their c
omponents (e.g, protein kinases and protein phosphatases) on signalling act
ivity. We argue that it is useful to view the signalling ecology as a vast
parallel distributed processing network of agents operating in heterogeneou
s microenvironments, and we conclude with an overview of the mathematical a
nd semantic methodologies that might help clarify this analogy between biol
ogical and computational systems. (C) 2000 Elsevier Science Ireland Ltd. Al
l rights reserved.