Cellular Automata Model for Proteomics and Its Application in Cancer Immunotherapy

摘要

This paper presents our first version of Protein modeling Cellular Automata Machine (PCAM). The peptide chain of amino acid backbone of a protein having n number of amino acids is designed with an 8n cell uniform CA employing one of the 64 three neighborhood CA (3NCA) rules. Each amino acid of a protein chain is modeled by a group of eight CA cells. Variation of the interaction pattern of a protein backbone under different physical conditions is modeled with different sixty-four 3NCA rules. Another set of twenty 8-bit patterns are next designed to encode the molecular structure of side chains of twenty amino acids. The eight CA cells representing an amino acid in the chain is initialized with the 8 bit pattern of its side-chain. A set of features extracted from evolution of PCAM are mapped to real life experimental results. The PCAM model is validated from cancer immunotherapy experimental results for MAb-PD-L1 interaction on multiple MAbs (Monoclonal Antibodies) with the protein PD-L1 associated in human immunity.