Lecturer: Prof. Uri Hershberg
Affiliation: School of Biomedical Engineering,
Science & Health Systems at Drexel University
Abstract:
We simulated the random flow of
information on real world biological
and social interaction networks and
multiple types of artificially
generated simulated networks. We
compared the pattern of movement
for that entities / units of
information. If motion was
completely unbiased movement
across hubs was the most frequent. If
on the other hand motion was
influenced by memory (i.e. the
direction of step n+1 was influenced
by the m steps before n) then motion
of information localized to
community structures in the
networks. This held true if
communities were uniform or of
uneven structure. Moreover, when
this method was applied to the
protein-protein interaction network
of c. elegans the communities /
modules we identified were clearly
of biological relevance.