compartmentalization of components within thecell, and with temporal profiles of activation, toobtain a more comprehensive picture of thefunctional organization of a cell. Still, distance inchemical space (i.e., number of links betweentwo distal nodes) is likely to be a major deter-minant of information processing that regulatesphenotypic behavior.The maps for individual ligands or cellularmachines show distinct patterns of motifs. Com-binations of ligands will likely produce manymore patterns of connectivity. Thus, a cellularsystem may not be a single network but rather anensemble of network configurations that areevoked by the stimuli-induced activation of var-ious parts of the system. Identifying these networkconfigurations and the functions they evoke islikely to provide more complete descriptions ofhow molecular interactions lead to cellularchoices between homeostasis and plasticity.References and Notes1. J. D. Jordan, E. M. Landau, R. Iyengar, Cell 103, 193(2000).2. U. S. Bhalla, R. Iyengar, Science 283, 381 (1999).3. U. S. Bhalla, P. T. Ram, R. Iyengar, Science 297, 1018(2002).4. R. Iyengar, Science 271, 461 (1996).5. R. D. Blitzer et al., Science 280, 1940 (1998).6. G. Lahav et al., Nat. Genet. 36, 147 (2004).7. D. Angeli, J. E. Ferrell, E. D. Sontag, Proc. Natl. Acad.Sci. U.S.A. 101, 1822 (2004).8. S. Mangan, A. Zaslaver, U. Alon, J. Mol. Biol. 334, 197(2003).9. S. Mangan, U. Alon, Proc. Natl. Acad. Sci. U.S.A. 100,11980 (2003).10. D. J. Watts, S. H. Strogatz, Nature 393, 440 (1998).11. L. A. Amaral et al., Proc. Natl. Acad. Sci. U.S.A. 97,11149 (2000).12. A. L. Barabasi, R. Albert, Science 286, 509 (1999).13. T. V. Bliss, G. L. Collingridge, Nature 361, 31 (1993).14. S. A. Siegelbaum, E. R. Kandel, Curr. Opin. Neurobiol.1, 113 (1991).15. Materials and methods are available as supportingmaterial on Science Online.16. O. Hvalby et al., Experientia 43, 599 (1987).17. H. Katsuki, Y. Izumi, C. F. Zorumski, J. Neurophysiol.77, 3013 (1997).18. H. Kang, E. M. Schuman, Science 267, 1658 (1995).19. N. Kashtan, S. Itzkovitz, R. Milo, U. Alon, Bioinfor-matics 20, 1746 (2004).20. P. V. Nguyen, T. Abel, E. R. Kandel, Science 265, 1104(1994).21. R. Bourtchuladze et al., Cell 79, 59 (1994).22. C. Pittenger, Y. Y. Huang, R. F. Paletzki et al., Neuron34, 447 (2002).23. G. Caldarelli et al., European Physical Journal B. 38,183 (2004).24. This research is supported by GM-54508 and GM-072853 from NIH and by an Advanced ResearchCenter grant from the New York State Office ofScience and Technology. A.M. is supported byPharmacological Sciences Training grant GM-62754.S.N. is the recipient of an individual predoctoralNational Research Service Award (GM-65065). R.D.B.is supported by National Institute on Drug Abusegrant DA15863. We thank B. Obrink, H. Dohlman, N.Hao, and J. Lisman for comments on the manuscript,M. Diverse-Pierluissi for help in identifying componentsof the secretory machine, S. Purushothaman for imple-menting the grid-coefficient function, and G. Kossinets(D. Watts laboratory, Columbia University, NY) for helpwith the initial analysis. Author contributions are de-scribed in the Supporting Online Material.Supporting Online Materialwww.sciencemag.org/cgi/content/full/309/5737/1078/DC1Materials and MethodsSOM TextFigs. S1 to S11Tables S1 to S3References20 December 2004; accepted 7 July 200510.1126/science.1108876Containing Pandemic Influenzaat the SourceIra M. Longini Jr.,1*Azhar Nizam,1Shufu Xu,1Kumnuan Ungchusak,2Wanna Hanshaoworakul,2Derek A. T. Cummings,3M. Elizabeth Halloran1Highly pathogenic avian influenza A (subtype H5N1) is threatening to cause ahuman pandemic of potentially devastating proportions. We used a stochasticinfluenza simulation model for rural Southeast Asia to investigate the ef-fectiveness of targeted antiviral prophylaxis, quarantine, and pre-vaccinationin containing an emerging influenza strain at the source. If the basicreproductive number (R0) was below 1.60, our simulations showed that aprepared response with targeted antivirals would have a high probability ofcontaining the disease. In that case, an antiviral agent stockpile on the orderof 100,000 to 1 million courses for treatment and prophylaxis would besufficient. If pre-vaccination occurred, then targeted antiviral prophylaxiscould be effective for containing strains with an R0as high as 2.1.Combinations of targeted antiviral prophylaxis, pre-vaccination, and quaran-tine could contain strains with an R0as high as 2.4.Theworldmaybeonthebrinkofaninfluenzapandemic (1–4). Avian influenza A (subtypeH5N1) is causing widespread outbreaks amongpoultry in Southeast (SE) Asia, with sporadictransmission from birds to humans (5)andlimited probable human-to-human transmission(6). Should an avian virus reassort with ahuman virus, such as influenza A subtypeH3N2, within a dually infected human host orreassort in a nonhuman mammalian species, orif mutation of the virus occurs, the resulting newvariant could be capable of sustained human-to-human transmission. The outbreak amonghumans would then spread worldwide via theglobal transportation network more rapidly thanadequate supplies of vaccine matched to thenew variant could be manufactured and distrib-uted (1, 7). The pressing public health questionsare whether and how we can contain the spreadof an emerging strain at the source or at leastslow the initial spread to give time for vaccinedevelopment. We used a discrete-time sto-chastic simulation model of influenza spreadwithin a structured geographically distributedpopulation of 500,000 people in SE Asia tocompare the effectiveness of various interven-tion strategies against a new strain of influenza.Here we examine the effectiveness of the tar-geted use of influenza antiviral agents (8 –12),quarantine, and pre-vaccination with a poorlymatched, low-efficacy vaccine in containing thespread of the disease at the source.We used information about rural SE Asia(13, 14) to construct the model population. Ourgoal was to represent the contact connectivityof a typical rural SE Asian population. Themodel population of 500,000 people wasdistributed across a space of 5625 km2, yieldinga density of 89/km2, which is approximatelythe population density of rural SE Asia (13).The 500,000 people were partitioned into 36geographic localities. This model is an ex-tension of a model used to simulate inter-ventions against pandemic influenza in theUnited States (12).The model Esee the supporting online ma-terial (SOM) for details^