Genome-Wide Prediction Of Cytokine-Receptor Interaction

Cytokines, 15～30KD, are a diverse group of cell intercellular messengers responsible for signaling variety of cell functions, such as immunity, hematopoiesis, chemotactic activities, cell maturation, proliferation, growth and differentiation through their interactions with respective receptors on cell membranes. The binding of cytokine to its complementary receptor is crucial for triggering cytokine-specific cellular activities. In this study, a series of Bioinformatics and statistical analyses were conducted to probe cytokine-specific activities by identifying cytokine-receptor interactions in the human genome. An advanced support vector machines system CytoSVM was constructed to successfully identify (estimated prediction accuracy 99.2%) 1,609 novel cytokine-receptor interactions from human genome. This covers 244 distinct novel cytokine receptors and 112 known cytokine receptors. To characterize the cytokine-receptor interactions, systematic and statistical analyses were demonstrated to illustrate their structural, transcriptional and genomic features. It is found that cytokine receptors intend to share some common signature patterns (domains/motifs) inside families. These patterns are conserved and specific in respective cytokine receptor families for execution and maintenance of their cytokine-receptor interactions. Data mining of the gene expression profiles revealed that only weak correlation between the expressions of cytokines and their complementary receptors. Chromosomal localization of cytokine receptors exhibits 11 obvious clusters of known and/or putative cytokine receptors distributed in human chromosomes; some of which are seem to be cellular activities-related. In particular, these newly identified interactions are involved in 31 characterized cellular activities in KEGG pathway database, 9 of these activities haven't been well studied or found for the known cytokines.