ESTIMATING ENZYME MELTING POINT BASED ON ENZYME AMINO ACID SEQUENCE USING MULTIPLE LINEAR REGRESSION

• Main Author: Dira Kurnia, Muhammad
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/81548
• Institution: Institut Teknologi Bandung

Enzymes are organic compounds that can accelerate biochemical reactions (catalysts). The structure of enzymes consists of two main parts: coenzymes that are active in the enzyme's catalytic function, and apoenzymes that encompass proteins in the form of amino acid chains or sequence. In industry, the use of enzymes is crucial for improving the quality of food products and cleaning agents. However, enzyme performance can be limited due to factors such as environmental temperatures that do not match the enzyme's characteristics. To address this issue, the development of identification or estimation systems for enzyme melting points or optimal temperatures based on physicochemical properties is necessary. Statistical methods such as Multiple Linear Regression can be used to predict enzyme melting points based on the physicochemical properties of amino acid sequences. Additionally, a method called Sequential Method with three different approaches can assist in evaluating significant physicochemical properties in estimating enzyme melting points. This method can be applied to data obtained from Novozymes, a global biotechnology company from Denmark, through Kaggle.com, which provides information about amino acid sequences and enzyme melting points. Based on experiments, preprocessed data or nummerized amino acid sequence based on physicochemical properties and subsequently, the sequential method yielded a multiple linear regression model with a combination of several predictor variables built from data without outlier residuals from the previous model.