Barrera Patiño Claudia P, Bonner Mitchell, Borsatto Andrew Ramos, Soares Jennifer M, Blanco Kate C, Bagnato Vanderlei S
Sao Carlos Institute of Physics (IFSC), University of Sao Paulo (USP), Sao Carlos 13566-590, SP, Brazil.
Biomedical Engineering, Texas A&M University, 400 Bizzell St, College Station, TX 77843, USA.
Antibiotics (Basel). 2025 Jul 20;14(7):729. doi: 10.3390/antibiotics14070729.
: In recent work, we have demonstrated that principal component analysis (PCA) and Fourier Transformation Infrared (FTIR) spectra are powerful tools for analyzing the changes in microorganisms at the biomolecular level to detect changes in bacteria with resistance to antibiotics. Here biochemical structural changes in were analyzed over exposure time with the goal of identifying trends inside the samples that have been exposed to antibiotics for increasing amounts of time and developed resistance. : All studied data was obtained from FTIR spectra of samples with induced antibiotic resistance to either Azithromycin, Oxacillin, or Trimethoprim/Sulfamethoxazole following the evolution of this development over four increasing antibiotic exposure periods. : The processing and data analysis with machine learning algorithms performed on this FTIR spectral database allowed for the identification of patterns across minimum inhibitory concentration (MIC) values associated with different exposure times and both clusters from hierarchical classification and PCA. : The results enable the observation of resistance development pathways for the sake of knowing the present stage of resistance of a bacterial sample. This is carried out via machine learning methods for the purpose of faster and more effective infection treatment in healthcare settings.
在最近的工作中，我们已经证明主成分分析（PCA）和傅里叶变换红外（FTIR）光谱是在生物分子水平分析微生物变化以检测对抗生素有抗性的细菌变化的有力工具。在此，分析了[具体对象]随暴露时间的生化结构变化，目的是确定在接触抗生素时间不断增加并产生抗性的样本中的变化趋势。：所有研究数据均来自对阿奇霉素、苯唑西林或甲氧苄啶/磺胺甲恶唑诱导产生抗生素抗性的样本的FTIR光谱，这些样本在四个不断增加的抗生素暴露时间段内呈现出这种抗性发展过程。：在此FTIR光谱数据库上使用机器学习算法进行的处理和数据分析，能够识别与不同暴露时间相关的最低抑菌浓度（MIC）值的模式，以及层次分类和主成分分析的聚类。：这些结果有助于观察抗性发展途径，以便了解细菌样本的当前抗性阶段。这是通过机器学习方法来实现的，目的是在医疗环境中更快、更有效地进行感染治疗。
