Machine learning helps to fully understand the principle of antibiotic sterilization

Xinhua News Agency, Beijing, May 11 - An international research team used machine learning algorithms to find that nucleotide metabolism plays an important role in the process of antibiotics killing bacteria. This will help people understand the principle of antibiotics more comprehensively and develop better sterilization methods based on this.

Researchers at institutions such as the Massachusetts Institute of Technology reported in the new US issue of Cell, that they use three antibiotics, ampicillin, ciprofloxacin, and gentamicin, and about 200 metabolically related substances. Combine them and observe the killing effect of different combinations on E. coli. They used machine learning algorithms to analyze the metabolic processes associated with bactericidal effects.

It was found that under the pressure of antibiotics, the supply of purine nucleotides used by bacteria for the synthesis of deoxyribonucleic acid (DNA) was insufficient, and production had to be intensified. This would consume a lot of energy and rapidly accumulate toxic metabolic waste, further making the bacteria survive. Deterioration, speeding up the death process.

Previous studies have shown that important pathways for antibiotic sterilization include key physiological processes that interfere with bacteria, such as DNA replication or cell wall construction. This study reveals the role of nucleotide metabolism, allowing people to more fully understand the sterilization principle of antibiotics.

It is not uncommon to use machine learning to process biological experimental data, but most models are "black box" type, knowing what experimental conditions correspond to what results, but do not know the principle. The new study adopted a "white box" model that combines experimental conditions, metabolic status, and bactericidal effects to determine the specific mechanisms by which drugs work.

Researchers say that the use of adjuvant drugs to enhance metabolic interference is expected to become a new means of killing bacteria and coping with bacterial resistance. The "white box" model can also be used to analyze the effects of different drugs on cancer, diabetes or neurodegenerative diseases. They are using similar methods to find out why tuberculosis has escaped antibiotic attacks and developed resistance.