DETERMINING RELATIVE IMPORTANCE OF HUMAN GUT MICROBIOTA, AGE, GENDER AND LIFESTYLE PATTERN AS A PREDICTOR FOR BMI USING LOGMPIE DATA

Abstract

We use the ‘Relative Abundance Table’ and ‘LogMPIE Study Metadata’ from the “Landscape of Gut Microbiome - Pan-India Exploration”, or LogMPIE dataset to find out the relative importance of human gut microbiota abundance (specifically genus), age, gender, and lifestyle pattern as a predictor for BMI (Body Mass Index). The LogMPIE data is taken from 1004 subjects and 993 unique microorganisms are reported along with BMI, age, and physical activity. We use Random Forest Regressor to find out the relative importance of the above-mentioned features (microorganism genus abundance, age, gender, and lifestyle pattern) in predicting the BMI of a subject. The objective here is not the prediction of BMI using the features but to find out the relative importance of these features as much as these affect the BMI.