Metagenomics is a derivation of the traditional branch of Genomics that studies the genome of communities of microorganisms extracted directly from their habitat, eliminating the need for laboratory cultivation. The reconstruction of the DNA strand of a genome is done through the sequencing process where smaller fragments of DNA are created, which will later be assembled into larger sequences in order to reach the original genome. For metagenomic data, this assembly process is more laborious, since several genomes of different organisms are present in the sample. Separating the sequences obtained in the assembly process into distinct categories according to the taxonomic level of organisms is a process called binning and is important to facilitate the subsequent analysis of the composition of the sample, for the discovery of phylogenetic relationships and new genes. This work seeks to compare the results of binning tools by defining a pipeline of metagenomic data analysis using two simulated datasets of 10 and 100 species of bacteria. Metaquast software was used to compare three assembly softwares (IDBA_UD, Megahit and MetasSPAdes) and two binning softwares (MetaBAT-2.12.1 and MaxBin-2.2.4). The quality of the bins generated by each tool was compared by constructing a Confusion Matrix and comparing the Accuracy, Sensitivity, Specificity and Accuracy metrics. As a result, MetaBAT excels at MaxBin in the quality of the bins generated in both sets of data, proving to be a good choice in performing binning of metagenomes.
|
