Principal Component Analysis and Clustering to reveal the conformation dynamics of EGFR with L858R and T790M Mutation

This paper investigates the dynamics of epidermal growth factor receptor EGFR (lung cancer related protein) with two most frequent mutations using principal component analysis and clustering. Mutation is the permanent alteration in the residue sequence, which in turns changes the protein structure and can lead to change the function of protein. Mutation of the epidermal growth factor receptor (EGFR) is a pathogenic factor in non-small cell lung cancer (NSCLC). Molecular Dynamics (MD) simulations enables us to study the motions of proteins and response of the protein to a particular drug. In this paper, we investigate behavior of EGFR (cancer related protein) before and after mutation. We used MD simulation by using the parallel computing CPUS/GPUS to generate 200ps simulation of 3 types of EGFR Protein. We analyze the motion of atomic trajectories with the help of principal component analysis and clustering. 20 principal components have been used in each EGFR mutant to describe the conformational dynamics of EGFR. The simulation result shows that resistive and mutated type EGFR captures the most variation in the first principal component followed by mutated type. The wildtype EGFR has the minimum variance in the first principal component. Based on this, we assume that the mutation destabilizes the overall structure of EGFR and it is difficult for a drug to bind the protein. This can be one of the reason for the drug resistance. We perform simple clustering in the principal component sub-space. Interesting clustering patterns are observed in each case. This study provides useful insight of EGFR mutant conformational dynamics and can be useful for drug discovery and computer aided targeted therapy in NSCLC treatment.