研究目的
To investigate GBS serotypes by incorporating machine learning techniques with MALDI-TOF MS to carry out the identification.
研究成果
The combination of MALDI-TOF MS and machine intelligence provides a practical means of clinical pathogen testing. The proposed models have been implemented in a web-based tool (GBSTyper) for efficient and effective detection of GBS serotypes.
研究不足
The performance of the predictive models declined in external validation, possibly due to the small sample size, especially for serotype Ia, Ib, and V. More samples are needed to train more robust ML models for clinical application.
1：Experimental Design and Method Selection:
The study was divided into three parts: data collection, data analysis, and prediction analysis. MALDI-TOF MS was used to obtain the mass spectra, and the samples were typed by geno-serotyping. Machine learning algorithms, such as SVM and RF, were used to construct predictive models for the five different serotypes.
2：Sample Selection and Data Sources:
A total of 787 GBS isolates were obtained from three research and teaching hospitals.
3：List of Experimental Equipment and Materials:
MALDI-TOF MS (Bruker Daltonik GmbH, Leipzig, Germany), BBL? Trypticase? Soy Agar with 5% Sheep Blood (TSA II) (Becton Dickinson, MD, USA), α-cyano-4-hydroxycinnamic acid (CHCA) matrix solution.
4：Experimental Procedures and Operational Workflow:
Fresh GBS colonies were smeared onto a MALDI target plate, overlaid with CHCA matrix solution, dried, and analyzed by MALDI-TOF MS. The mass spectra were acquired in the mass range of 2000 to 20,000 m/z in linear mode.
5：Data Analysis Methods:
The binning method was used to extract features from the mass spectra. Feature selection was conducted by OneR and PCC. Predictive models were trained and validated using 5-fold cross-validation and an independent testing dataset.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
