Um modelo de deep learning para classificação de organismos vivos utilizando o segmento 18S rRNA da sequência genética

Abstract

Taxonomy, in the field of biology, is the science that classifies living beings hierarchically according to common characteristics. Currently, there are several techniques for the taxonomic classification of organisms through genetic sequencing, commonly using metagenomic analysis. However, these techniques tend to be computationally expensive and, in scenarios where there are mutations or genetic variations within the same subgroup, they may fail or be inconclusive. With the recent advances in the field of artificial intelligence and machine learning, new potential forms of classification are being studied and developed. In this work, a taxonomic classifier solution based on a deep learning model is presented and a comparative analysis is also made between the results of the presented solution and the results obtained with the q2-feature-classifier of the QIIME2 platform. The results obtained show that the developed solution achieves greater accuracies, especially at the Species level