Mykrobe Predictor Software Identifies Antibiotic Resistance
Researchers from the Wellcome Trust Centre for Human Genetics have developed a new software that predicts antibiotic resistance in bacteria by analyzing their genome. The new tool, called Mykrobe Predictor, was tested in Staphylococcus aureus and Mycobacterium tuberculosis, and the results were similar to common phenotypic tests. The study has been published in Nature Communications.
The increase in bacterial antibiotic resistance has prompted scientists to try to find fast and cheap techniques to identify drug resistant bacteria in clinical settings. Lately, DNA sequencing has become more accessible than ever, convincing researchers to consider this technology a potential tool to detect antibiotic resistance. DNA sequence can provide much more information than phenotypic tests, where bacteria are cultured in media supplemented with different antibiotics. Sequencing a pathogen’s genome can inform about the species, the strain and the mutations that make it resistant.
Resistance prediction in just hours
The new software tool Mykrobe Predictor recognized different bacterial species in mixed samples, like S. aureus and M. tuberculosis. 99% of S. aureus samples where identified correctly, a number comparable to phenotypic testing. The software can find antibiotic resistance genes from raw DNA data in just three minutes.
For M. tuberculosis, however, there was a 15% rate of false negatives. This result can be attributed to the incomplete knowledge of the genetic mechanisms for resistance. On the plus side, the software can be quickly updated when new research results in antibiotic resistance are obtained.
Mykrobe Predictor was tested in three hospitals in the United Kingdom. Data generation took 16 hours of sequencing on the Illumina MiSeq, and just seven hours on a MinION. Both devices are extremely faster than the phenotypic method, which can take weeks with species like M. tuberculosis.
The bioinformaticians that created Mykrobe Predictor have released the code for free on GitHub. The software is easy to use, and creates reports about infection susceptibility and resistance.