Networks' characteristics are important for systems biology

• DOI: 10.1017/nws.2014.13
• Publication Date: 2014-09-03T08:58:14Z
• Authors: ANDREW K. RIDER; TIJANA MILENKOVIĆ; GEOFFREY H. SIWO; RICHARD S. PINAPATI; SCOTT J. EMRICH; MICHAEL T. FERDIG; NITESH V. CHAWLA
• Keywords: Network motif; Biological data; Gene regulatory network; Network Analysis

A fundamental goal of systems biology is to create models that describe relationships between biological components. Networks are an increasingly popular approach this problem. However, a scientist interested in modeling (e.g., gene expression) data as network quickly confounded by the problem: how construct network? It fairly easy network, but it for problem being considered? This important with three issues: How weight edges order capture actual interactions? What effect type experiment used collect from which constructed? prune weighted (or what cut-off apply)? Differences construction networks could lead different interpretations. Indeed, we find there statistically significant dissimilarities functional content and topology co-expression constructed using edge weighting methods, types, cut-offs. We show types known interactions, such those found through Affinity Capture-Luminescence or Synthetic Lethality experiments, appear significantly varying amounts ways. Hence, demonstrate questions may be answered networks. Consequently, posit taken build can matched get targeted answers. More study required understand implications inference approaches draw reliable conclusions field biology.