Skip to main content
Fig. 2 | Microbiome

Fig. 2

From: HMD-ARG: hierarchical multi-task deep learning for annotating antibiotic resistance genes

Fig. 2

HMD-ARG database composition and the HMD-ARG database construction pipeline. a The statistics of the HMD-ARG database. The number of sequences belonging to each antibiotic family is different. Meanwhile, different genes can resist the same drug with different mechanisms, which are shown in different colors. b Different databases have various numbers of sequences as well as different labeling information. HMD-ARG database is currently the largest one. At the same time, it is the most comprehensive one, with resistant antibiotic class, resistant mechanism, and gene mobility labeled. c To construct the database, we merged the sequences from seven existing databases, followed by a post-processing step to remove duplicates. Then, we either used the existing tools or manual curation to assign the annotations from three aspects to each sequence in the database

Back to article page