TaxHorn: A Novel, Taxonomy-Aware, Beta-Diversity Metric for Comparison of Microbiome Populations

Overview

While there are numerous metrics currently available for measuring beta diversity, none of these are 108 particularly optimized for work within the human microbiome. I set out to create a new distance metric, optimized for the microbiome with respect to taxonomy. T o determine what would be required of such a tool, I first tested four existing metrics (Bray-Curtis, Horn-Morisita, Euclidean, and Unifrac) against a microbiome dataset containing samples of microbial populations from several individuals of varying degrees of health. I generated every possible pairwise combination of these samples, to which the distance metrics were then applied to. The information gained from this showed me what areas needed to be improved, and where I should begin. Of all of the metrics, the Horn-Morisita index proved most useful. I then designed a custom metric, which calculates said index to each taxonomic level of every sample. Finally, I applied a weighted mean to merge the seven remaining diversities into a single value for each sample with less focus on the immense variety of species. This metric, taxHorn, allows one to accurately estimate the degree to which two microbiome samples are different. This could potentially be used to detect the beginning of bacterial infections within immunocompromised individuals so that they can be medicated before an infection can fully develop and to further our understanding of the microbiome as a whole.

Competition history

• JSHS 2023 Category not listed

Resources

• National JSHS abstract book ↗

Related projects

Closest projects by meaning, across every fair and year in the corpus.