Eigen RadarScience
Archive
Analysis

MAC-Explorer gives 2,737 bacterial genomes a common classification

MAC-Explorer brings a shared genome-based classification to a bacterial group linked to infections in people and animals. A peer-reviewed analysis of 2,737 genomes identified 15 species-level clusters and tested an identification tool on independent clinical bacterial genomes. The work addresses inconsistent laboratory labels; its classification performance does not establish better treatment outcomes, and rare lineages need broader evaluation.

Science··Night
Closed genome-sequencing instrument seen through a laboratory doorway.

Bacterial genomes receive shared labels

MAC-Explorer, a web tool for identifying members of the Mycobacterium avium complex, implements a shared classification for bacteria associated with infections in humans and animals. The peer-reviewed study compared 2,737 high-quality bacterial genomes after removing duplicate sequences and applying quality filters. Inconsistent names for closely related bacteria have complicated comparisons between laboratories and studies. The new framework assigns genomes to common groups using their genetic similarity.[1], [2]

Researchers used two methods to calculate average nucleotide identity, a measure of similarity between genome sequences. Approximate thresholds of 95 per cent for species and 98 per cent for subspecies gave an operational basis for grouping. The consensus identified 15 species-level clusters. Some existing names fell within the same genomic group, while other groups retained finer lineage structure. Provisional labels for genomic groups carry no formal standing as newly described species.[1]

An identification tool reproduces the genomic groups

The identification component uses random-forest classifiers, which combine multiple decision trees. They learn from the frequencies of short DNA sequences to reproduce the labels established by genome comparison. This places the classification in a practical tool rather than requiring every user to rebuild the comparative analysis. Its performance measures agreement with those reference labels. Because training labels come from the same genomic framework, classifier accuracy supplies evidence for implementing that framework, rather than independent proof that the taxonomy is correct.[1]

Independent isolates test the tool’s portability

An external evaluation used 74 clinical bacterial genomes that had been excluded from training. All species and subspecies assignments matched the genome-based reference strategy; finer lineage assignments agreed for 71 genomes. These were tests of bacterial identification using independent isolates. Patient outcomes and the success of treatment decisions were outside the evaluation.[1]

Public genomes underrepresent environmental, veterinary and other non-human sources. Several rare groups contain few examples, so broader evaluation across countries, laboratories and sequencing workflows remains necessary. Routine clinical adoption also depends on access to whole-genome sequencing and appropriate regulatory validation. These constraints shape how widely a common classification can be used in laboratory reporting.[1]

References

  1. News sourceGenome MedicineAnalysis of 2,737 bacterial genomes harmonized MAC classification↩1↩2↩3↩4↩5
  2. News sourceScienmagAnalysis of 2,737 bacterial genomes harmonized MAC classification↩
Space and explorationEnergy transition and climate

More analysis from this desk

View all Brief analyses from the Science desk →