Important protein activities can affect the cellular life cycle and human health. The occurrence and progression of various diseases are linked to the abnormal expression of specific proteins; therefore, it is feasible to develop drugs that normalize the behavior of the target protein to treat such diseases.

However, occasionally the protein mutates. As a result, its properties may alter, and the drugs' efficacy will diminish. One possibility is to analyze the structure of the proteins to address this issue, which can provide crucial information about the behavior-related properties.

Cancer is one of the most dangerous diseases, and lung cancer is the deadliest. Non-small cell lung cancer (NSCLC) is the most common form of lung cancer. The epidermal growth factor receptor (EGFR) and its signaling pathways are known to play a significant role in lung cell proliferation. The dimerization of EGFR family members and other receptor tyrosine kinases (RTKs) is a crucial regulator of lung cell life cycle signals. Mutations in the kinase domain of EGFR may disturb signaling networks and result in cancer. In addition, several generations of EGFR drugs exhibit drug resistance due to genetic mutations, necessitating research into potential pathogenic mechanisms.

First, we investigated the correlation between EGFR mutations and EGFR-RTK crosstalk in the signaling network to identify the EGFR mutation-induced drug resistance mechanism. Using computational methods based on molecular dynamics (MD) simulations, the wild-type (WT) EGFR and several of its mutants are used to measure the EGFR-RTK interactions. We developed new geometrical methods based on alpha shape modeling, including the matching rate of atomic solid angles in interfaces and center-of-mass distances between interacting atoms. In addition, conventional estimations of the free energy of binding are used as a standard. The drug-sensitive EGFR mutant exhibited a looser EGFR-RTK crosstalk, while the drug-resistant EGFR mutant exhibited a tighter crosstalk. It is a potential mechanism for drug resistance that the EGFR-RTK crosstalk is amplified caused by EGFR mutations.

Then, we investigate the effect of mutations in the EGFR dimers themselves when induced by drugs. We created a variety of EGFR mutants capable of forming both homo-dimer and hetero-dimer complexes. MD simulation is performed to investigate the geometrical properties of structures. To characterize the intensity of binding, geometric properties were calculated. A Wilcoxon rank-sum test was applied based on extracted properties to determine the differences between mutations. In all protein-drug systems, drug-sensitive mutants have tighter interactions with the corresponding RTK in the complex, whereas drug-resistant mutants are loosely bound. This work suggests an additional possible cause for drug resistance.

In addition, we attempt to use geometrical tools to examine the changes in geometrical properties caused by protein mutations. The surface atoms of EGFR and its mutations are extracted, and their activities are characterized based on their average moving distance and mass. The results indicate that mutations on the kinase domain resulted in changes to the geometrical properties and the distribution of active sites. Moreover, there are observable distinctions between drug-resistant and drug-susceptible mutations. Changes in geometrical properties affect the specific antigenicity of drugs and decrease their binding affinity. By providing alternative tools to detect new binding sites and contributing to the design of drugs, the methods help to explain drug resistance on a geometrical level.

The emergence of the COVID-19 pandemic necessitates research into its mechanism and treatment development. Using our structural analysis technique, we examined the theoretical efficacy of a potential drug, dexamethasone. The binding affinity of dexamethasone to the SARS-CoV-2 protease was used as a criterion for evaluation, and several ideal and potential inhibitors were collected for comparison, docking drug molecules to target proteins and simulating the system with MD. According to geometrical analysis results, dexamethasone behaved similarly to other inhibitors on these indicators. Therefore, the efficacy of dexamethasone as a treatment may be due to its glucocorticoid and potent inhibitor properties.

Covid-19 protein mutations can also cause immune evasion and diminish the efficacy of vaccines and antibodies. To explain the mechanism, the geometrical properties of the receptor-binding domain in the SARS-CoV-2 spike protein are examined. Several significant variants serve as examples, while the wild-type model serves as a guide, utilizing the average distance traveled by surface atoms to quantify their activity. The distribution of active sites on the surface of a protein is correlated with its specific antigenicity, and mutations can alter the binding abilities of drugs and antibodies, resulting in immune evasion.

This thesis develops a framework based on geometrical properties to explain the mechanism of drug resistance. As a result, it will lead to a greater comprehension of the mechanisms underlying EGFR mutation-induced drug resistance and promote the design of innovative drugs. Furthermore, as the extracted structural properties aid in understanding the drug response mechanism at the atomic level, the structural analysis tools developed in this study can be used in other contexts to evaluate the efficacy of drugs for various diseases.