基于机器学习的甲状腺癌患者肠道微生态特征分析OA

To identify diagnostic microbial biomarkers,machine learning algorithms were used to analyze the gut microbiome profiles of patients with thyroid cancer.Gut microbiome sequencing data of thyroid cancer were obtained from the National Center for Biotechnology Information(NCBI)database.The Kraken2 tool was used for taxonomic annotation and construction of an operational taxonomic unit(OTU)abundance table.The H2O automated machine learning platform was applied,integrating algorithms such as generalized linear model(GLM),distributed random forest(DRF),and deep learning to screen for key characteristic microorganisms,combined with the Wilcoxon rank-sum test to analyze abundance differences between groups.Through cross-validation across multiple models,key OTUs including Sutterella,Emergencia,Lactococcus,and Carnobacterium were identified.Differential analysis showed that,compared with the healthy control group,the abundance of Sutterella in the gut of thyroid cancer patients was significantly decreased(P＜0.001),while the abundances of Emergencia,Lactococcus,and Carnobacterium were significantly increased(P＜0.001).These results indicate that specific alterations exist in the gut microbiota structure of thyroid cancer patients.The identified microorganisms may serve as auxiliary diagnostic biomarkers and novel targets for research on thyroid-gut axis.