Journal of global antimicrobial resistance · 2026
A metatranscriptome based approach to predict multidrug resistance phenotypes in Klebsiella pneumoniae without culturing.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Xiang Zhoufu, Guan Jiaojiao, Wang Han, Wu Wanhua, Chen Sijun, Wang Siqi, Jiang Zihao, Hu Yiyi, Xiang Miao, Huang Tingsheng, Tian Hongpan, Sun Yanni, Li Yirong, Hou Wei, Chen Liangjun
Аннотация
INTRODUCTION: Antimicrobial resistance (AMR), particularly multidrug resistance (MDR) in Klebsiella pneumoniae, has emerged as a significant global public health challenge. However, accurate resistance phenotype prediction remains difficult.
METHODS: Metatranscriptomic analysis was performed on bronchoalveolar lavage fluid from 31 patients. After removing host and rRNA sequences, antimicrobial resistance genes (ARGs) were identified using the CARD database. ARG expression profiles trained a machine learning model for resistance prediction. Performance was compared to conventional methods (culture, susceptibility testing, whole genome sequencing).
RESULTS: Our methodology successfully captured K. pneumoniae in all 31 clinical samples, achieving 100% accuracy. In contrast, conventional methods resulted in two misidentifications, with an accuracy of 93.5%. Furthermore, for resistance phenotype prediction, our machine learning-based model achieved strong results: an AUROC of 0.86 and an F1 score of 0.77 for Carbapenems, and an AUROC of 0.81 and an F1 score of 0.69 for Aminoglycosides. Importantly, the workflow reduced the diagnostic time to 15 h, compared to the traditional culture-based methods, which require at least 29 h.
CONCLUSION: These advancements of our approach establish a novel paradigm for rapid AMR detection and position this method as a valuable supplementary tool to support clinical AMR diagnosis and precision antimicrobial stewardship against multidrug-resistant pathogens.
METHODS: Metatranscriptomic analysis was performed on bronchoalveolar lavage fluid from 31 patients. After removing host and rRNA sequences, antimicrobial resistance genes (ARGs) were identified using the CARD database. ARG expression profiles trained a machine learning model for resistance prediction. Performance was compared to conventional methods (culture, susceptibility testing, whole genome sequencing).
RESULTS: Our methodology successfully captured K. pneumoniae in all 31 clinical samples, achieving 100% accuracy. In contrast, conventional methods resulted in two misidentifications, with an accuracy of 93.5%. Furthermore, for resistance phenotype prediction, our machine learning-based model achieved strong results: an AUROC of 0.86 and an F1 score of 0.77 for Carbapenems, and an AUROC of 0.81 and an F1 score of 0.69 for Aminoglycosides. Importantly, the workflow reduced the diagnostic time to 15 h, compared to the traditional culture-based methods, which require at least 29 h.
CONCLUSION: These advancements of our approach establish a novel paradigm for rapid AMR detection and position this method as a valuable supplementary tool to support clinical AMR diagnosis and precision antimicrobial stewardship against multidrug-resistant pathogens.