The emerging chemical patterns applied in predicting human toll-like receptor 8 agonists

Abstract

Toll-like receptors (TLRs) are important pattern recognition receptors to human innate immunity, which can recognize pathogen-associated molecular patterns and initiate innate immune responses. As the receptor of single stranded RNA (ssRNA), toll-like receptor 8 (TLR8) has potential in the treatment of tumors, microbial infection, and inflammatory diseases. Herein, an emerging chemical pattern (ECP) method was utilized to predict the key chemical patterns of TLR8 agonists. Based on the ECPs discovered, a robust and predictive ECP model was derived with prediction accuracies of 83.3%, 81.0%, and 80.0% for 132 training samples, 79 validation samples, and 75 test samples, respectively. When the ECP model was applied with a molecular docking method, the hit rate of TLR8 agonists was greatly enhanced. The results of ECP-based hierarchical cluster analysis and Connolly surface analysis of the TLR8 receptor showed that the H-bonding, hydrophilic and hydrophobic potentials as well as the unbalanced degree of property distributions are very important for distinguishing the TLR8 agonists from non-agonists.