Liu Ying, Wang Xudong, Wu Liangquan, Shi Zhenzhen, Zhu Miao, Ye Linhui, Xu Ping
Department of Tuberculosis, The Second Hospital of Nanjing, Nanjing University of Chinese Medicine, Nanjing, China.
Department of Critical Care Medicine, the Affiliated Xuzhou Municipal Hospital of Xuzhou Medical University, Xuzhou, China.
Microbiol Spectr. 2025 Sep 2;13(9):e0006525. doi: 10.1128/spectrum.00065-25. Epub 2025 Aug 5.
The phenomenon of antimicrobial resistance (AMR) often results in treatment failure and restrictions on precision medicine, emphasizing the need for molecular diagnosis of drug resistance. The current use of machine learning (ML) techniques based on whole genome sequencing (WGS) data offers a more precise prediction of phenotypes. We incorporated WGS data from 3979 strains in our study. We modeled 10 common antibiotics using three types of features: gene, single nucleotide polymorphism (SNP), and k-mer to identify the best model and to determine which feature values most significantly contributed to the model's performance. The area under the curve (AUC) values of 40 mL models for 10 antibiotics ranged from 0.8345 to 0.9995. We noted that the performance indices such as the AUC of the gene model (0.9311-0.9992) and the integrated model (0.9313-0.9995) were markedly better than the SNP model (0.8345-0.9933) and the k-mer model (0.9024-0.9969). The best model AUC values for six antibiotics-cefoxitin, tetracycline, methicillin, gentamicin, erythromycin, and clindamycin-were over 0.99; nine antibiotic models had AUC values over 0.96, and all could effectively predict AMR phenotypes. Additionally, we discovered that certain non-AMR genes, such as the X998_03220 gene, significantly contributed to drug resistance prediction and overlapped in various antibiotic-related models simultaneously. Our study developed ML models that can reliably predict AMR phenotypes for commonly used antibiotics in . We also identified potential molecular markers that can contribute to precision medicine implementation and healthcare cost reduction.
In our study, we developed a machine learning (ML) model that reliably predicts the antimicrobial resistance (AMR) phenotypes of to commonly used antibiotics. This model not only predicts AMR phenotypes but also identifies potential molecular markers, which could facilitate the implementation of precision medicine and contribute to reducing healthcare costs. The integration of diverse biomarker types is crucial for enhancing model performance; however, their effectiveness may vary depending on the specific antibiotic in question. Furthermore, our pan-genome-based characterization has revealed novel potential molecular markers associated with AMR, thereby enhancing our comprehension of the underlying molecular mechanisms of AMR in . The expedited implementation of early and targeted antimicrobial therapies for infections is essential for advancing precision medicine and can potentially lead to significant healthcare cost savings.
抗菌药物耐药性（AMR）现象常常导致治疗失败，并限制精准医学的应用，这凸显了对抗菌药物耐药性进行分子诊断的必要性。目前基于全基因组测序（WGS）数据使用机器学习（ML）技术能够更精确地预测表型。我们在研究中纳入了3979株菌株的WGS数据。我们使用三种类型的特征——基因、单核苷酸多态性（SNP）和k-mer，对10种常用抗生素进行建模，以识别最佳模型，并确定哪些特征值对模型性能的贡献最为显著。针对10种抗生素的40个模型的曲线下面积（AUC）值在0.8345至0.9995之间。我们注意到，基因模型（0.9311 - 0.9992）和整合模型（0.9313 - 0.9995）的性能指标，如AUC，明显优于SNP模型（0.8345 - 0.9933）和k-mer模型（0.9024 - 0.9969）。六种抗生素（头孢西丁、四环素、甲氧西林、庆大霉素、红霉素和克林霉素）的最佳模型AUC值超过0.99；九个抗生素模型的AUC值超过0.96，并且所有模型都能有效预测AMR表型。此外，我们发现某些非AMR基因，如X998_03220基因，对耐药性预测有显著贡献，并且在各种与抗生素相关的模型中同时存在重叠。我们的研究开发了能够可靠预测常用抗生素AMR表型的ML模型。我们还确定了有助于实施精准医学和降低医疗成本的潜在分子标记。
在我们的研究中，我们开发了一种机器学习（ML）模型，该模型能够可靠地预测对常用抗生素的抗菌药物耐药性（AMR）表型。该模型不仅能预测AMR表型，还能识别潜在的分子标记，这有助于精准医学的实施并有助于降低医疗成本。整合多种生物标志物类型对于提高模型性能至关重要；然而，它们的有效性可能因所涉及的特定抗生素而异。此外，我们基于泛基因组的特征分析揭示了与AMR相关的新型潜在分子标记，从而增强了我们对AMR潜在分子机制的理解。对于感染尽快实施早期和有针对性的抗菌治疗对于推进精准医学至关重要，并且有可能大幅节省医疗成本。
