Performance evaluation of molecular markers for the taxonomic classification of the clade Bacillus subtilis

PETITTI, T.; TORRES MANNO, MARIANO A.; FIGUEROA, SOL A.; DAURELIO, LUCAS D.; ESPARIZ, MARTIN

Cordoba

Congreso; XVIII Congreso Argentino de Microbiología Genera (SAMIGE); 2022

SAMIGE

The genus Bacillus is formed by multiple clades, among them, the Bacillus subtilis clade stands out. The Bacillus subtilis clade is usually classified in 5 groups, B. subtilis group, B. amyloliquefaciens group, B. gobiensis group, B. pumilus group, and B. licheniformis group. Nevertheless, due to the high diversity of the clade, the taxonomic classification is challenging, having reported multiple cases of incorrect classifications or incoherences between taxonomic classification and genomic or phenotypic characteristics. However, the proper assignment is critical since these assignments are used to estimate the performance and safety of bacteria, which has an impact on their use in industry and agriculture.In order to verify and correct the taxonomic classification of genomes belonging to the Bacillus subtilis group, and thus generate a curated database, 2400 sequences belonging to the five groups were downloaded from GenBank. These genomes and 225 genomes provided by ARS-USDA were quality filtered, eliminating 133 sequences. The species-level taxonomic identity of the 2492 remaining strains was validated or reassigned using Average Nucleotide Identity (ANI) and multi-locus sequence analysis (MLSA). Thus, 29.5% of the sequences were reassigned. In turn, 148 strains were classified as new genomospecies, named genomospecies 1 to 12. After that, the performance of the 16S gene to classify strain groups was tested. An accuracy of 0.78 and a recall of 0.81 were observed for all groups. Moreover, B. licheniformis group shows an accuracy and recall as low as 0.37 and 0.38, respectively. Hence, a new gene marker is needed to classify groups, and therefore, species of Bacillus subtilis clade. With that attempt, the performance of 14 genetic markers was evaluated. The accuracy and recall to assign species before correcting the taxonomic assignments was 0.71 and 0.26 respectively. While, with the curated database, accuracies and recalls to classify clades were 1 for both cases. On the other hand, these parameters had values from 0.85 to 1 and a recall from 0.95 to 1 to classify species. Based on these results, to assign a clade we propose a shared ID% cutoff of 86.89 with the sequence of the highest accuracy marker. In sum, we have selected genetic markers that have an accuracy similar to that obtained through whole genome analyses. This would allow assignments of new isolates of the group with good precision and faster than actual methods and is the first step to develop methods for mass assignments in metagenomic analyses.