Deep learning-based prediction of chemical accumulation in a pathogenic mycobacterium
2024

Deep learning predicts how drugs accumulate in a harmful bacterium

Sample size: 1528 publication 10 minutes Evidence: high

Author Information

Author(s): Sullivan Mark R., Rubin Eric J.

Primary Institution: Harvard T.H. Chan School of Public Health

Hypothesis

Can deep learning models accurately predict drug accumulation in Mycobacterium abscessus based on chemical structure?

Conclusion

Deep learning models can effectively predict drug accumulation in Mycobacterium abscessus, which may help in developing more effective antibiotics.

Supporting Evidence

  • Deep learning models showed high accuracy in predicting drug accumulation.
  • Two antibiotics, clofazimine and bedaquiline, were identified as having high accumulation in M. abscessus.
  • Predicted accumulation correlated with the efficacy of compounds against M. abscessus.

Takeaway

Scientists used computers to figure out how well different medicines can get into a tricky germ. This can help make better medicines to fight that germ.

Methodology

Liquid chromatography-mass spectrometry was used to measure the accumulation of 1528 approved drugs in Mycobacterium abscessus, followed by training deep learning models to predict drug accumulation based on chemical structure.

Potential Biases

Potential bias in the training data due to the limited diversity of the drug library used.

Limitations

The study's predictive models may not generalize to all bacterial species due to unique accumulation mechanisms.

Statistical Information

P-Value

p<0.05

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1101/2024.12.15.628588

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication