Community analysis in interactomic regulation networks

In order to investigate interactomic networks, in fact, a simple but powerful idea, as proved by Google usefulness and consequential success, is to investigate the ranking of each actor relationships to each other, being such actor either an internet page, as in the original Page algorithm, or a gene - or codified protein - in our case. It is worth noticing, as pointed out in a seminal review by Vidyiasagar, that the recent randomized approach introduced by Tempo and coworkers could be the technological key to drastically reduce, at the cheap price of a limited loss of precision, the over-helming computational complexity that would prevent to apply Page ranking to the analysis of every significant interactomic network, besides the almost-toy sub-networks already investigated, as for instance by Zaki In this paper, we would like to investigate a possible complementary approach, recently proposed by Landi & Piccardi, to our interactomic regulation network. As a public available benchmark, the data used in the reported work by Zaki are also used, in order to investigate which features of our proposed approach are possibly useful as a complement to even improve the already powerful Page approach on one side, or eventually able to surrogate it in a less costly way, even taking into account the recalled randomized economy.