Currently, bioinformatic tools focused on predicting protein structures and simulating the
interactions between these proteins and different molecules are being developed. Among them, the AlphaFold program stands out as an innovative artificial intelligence system used to predict protein structures. However, the application of these computational protocols is still undergoing validation. Specifically, a general protocol that allows the routine use of these tools in drug design or personalized medicine has not yet been found. The objective of this work is primarily to determine the validity of the AlphaFold method for predicting the structures of proteins that are therapeutic targets against cancer. Firstly, it has been determined that the structures predicted by AlphaFold are highly similar to the experimental ones in terms of global structure, indicating its potential and the projection that this system has. However, a more detailed analysis of these structures generated with AlphaFold has revealed that there are variations compared to experimental structures in certain domains that are generally disordered but crucial for the physiological functioning of the protein. In fact, they play a crucial role in the regulation and transmission of signaling pathways and, especially, they affect kinases due to their inherent nature. Subsequently, the interaction of these predicted proteins with different molecules, with which they have a certain affinity, has been studied using molecular docking programs. It has been observed that, overall, the AlphaFold structures do not provide a better prediction of the receptorligand interaction than the experimental structures. Finally, it has been demonstrated that the structures predicted by AlphaFold, relaxed through molecular dynamics, do not improve the outcome of the initial prediction. This study has shown that bioinformatic tools are in development and could be very useful for research. However, further work is needed to improve these tools in order to better analyze and understand the complex interactions that occur at the physiological level. This advancement would drive the discovery and development of new therapies targeting proteins responsible for diseases such as cancer
DUGiDocs