Review and Comparison of Antimicrobial Resistance Gene Databases

Abstract

As the prevalence of antimicrobial resistance genes is increasing in microbes, we are facing the return of the pre-antibiotic era. Consecutively, the number of studies concerning antibiotic resistance and its spread in the environment is rapidly growing. Next generation sequencing technologies are widespread used in many areas of biological research and antibiotic resistance is no exception. For the rapid annotation of whole genome sequencing and metagenomic results considering antibiotic resistance, several tools and data resources were developed. These databases, however, can differ fundamentally in the number and type of genes and resistance determinants they comprise. Furthermore, the annotation structure and metadata stored in these resources can also contribute to their differences. Several previous reviews were published on the tools and databases of resistance gene annotation; however, to our knowledge, no previous review focused solely and in depth on the differences in the databases. In this review, we compare the most well-known and widely used antibiotic resistance gene databases based on their structure and content. We believe that this knowledge is fundamental for selecting the most appropriate database for a research question and for the development of new tools and resources of resistance gene annotation.

Keywords: annotation of antimicrobial resistance genes; antimicrobial resistance gene database; antimicrobial resistance genes.

Figure 1

ARG and sequence content of the databases. Only antibiotic and biocide resistance genes were considered for the plot. For each database on the x axis, the number of unique sequences and the corresponding number of unique genes were determined. The y axis represents the number of genes and sequences. Red bars show the gene number while blue bars represent the number of sequences stored in each database.

Figure 2

Number of unique genes for each antibiotic class stored in CARD and ResFinder. Bars represent the number of genes in each unique antibiotic or biocide categories, where colors are associated with the specific antibiotics themselves. As one gene can confer resistance to multiple antibiotics, it is possible that the same gene is counted for two or more antibiotics. The plots show the data for CARD (A) and ResFinder (B), respectively.

Figure 3

Number of genes conferring resistance through mutations for each microbial genus in CARD, NDARO and MEGARes databases. Genes conferring resistance through mutations was calculated at the genus level for each microbe stored in each database. Only one group could not be summarized at the genus level (propionibacteria). Microbial genus is on the x axis and the number of genes associated with each group in the database is represented by the y axis. Rows show the data separately for each database. Columns are colored by the microbial genus.