Large-scale gene coexpression is a powerful tool to elucidate gene functional modules. Over the past 10 years, gene coexpression has been extensively used, especially in plant science, for the prioritization of genes with a particular function of interest (1–5). In addition to microarray technologies, recent RNAseq data produced by high-throughput sequencing technologies are available for the construction of coexpression systems (6–8). The accumulation of RNAseq data opens the possibility to apply the coexpression approach even to non-model species. In animal science, coexpression information is largely used as part of the supporting information to predict protein–protein interactions, such as with the STRING (9), IMP (10) and Funcoup (11) databases. A limited number of databases are now focusing on coexpression information as the main content, such as GeneFriends (12), GeneMANIA (13) and STARNET2 (14). Gene coexpression has a long range to search, even for weak functional associations. Thousands of genes are meaningfully coexpressed for one cellular function (15). However, this long-range characteristic becomes problematic when a researcher wants to search for only directly associated genes, as in protein–protein interaction relationships. To focus
