MANIPULATING PROTEIN SOLUBILITY BASED ON MACHINE LEARNING PREDICTIVE MODEL

Abstract

Proteins are complex and big molecules that play a number of critical roles in the human body. They are required for the formation, function, and regulation of the cells that make up, and they primarily act in cells. The amino acid sequence of a protein determines its biological function. Denaturation is the process through which a protein can be unfolded. Temperature, activators, pH levels, and inhibitors are just a few of the things that might influence enzyme activity. That's a nice one, temperature. Proteins alter form as a result of temperature variations. After being created in response to external instructions, most proteins can be controlled by covalent changes or interactions with other molecules. Changing the rate of protein breakdown can also be used to manipulate protein levels within cells. Proteins have a crucial role as enzymes, which are required to catalyze nearly all biological activity. As a result, modulating enzyme activity is crucial for influencing cell behavior. The ability of recombinant proteins to operate as biocatalysts is determined by their activity. High-activity protein biocatalysts are less expensive. Because it assists in protein testing and optimization, this thesis describes a model that predicts protein activity based on amino acid sequence. However, due to limited data on protein activity, such models can no longer be developed. Because the activity and solubility of some proteins are connected, publically available solubility data might be utilized to construct models that predict protein solubility based on the sequence. Based on the sequence, the models might be used to estimate protein activity. Despite a large research on predicting protein solubility from sequence, the solubility was predicted using a machine learning model, which will aid the studies greatly. And we can greatly increase the solubility of proteins in vivo with the help of machine learning models.