矮身材儿童差异表达肠道菌群分析

Analysis of differentially expressed gut microbiota in short-stature children

目的分析矮身材儿童与正常身高儿童的肠道菌群表达差异,探索肠道菌群变化与矮身材的关系。方法选择矮身材患儿66例纳入矮身材组,56例健康儿童纳入对照组,分析并比较两组肠道菌群的多样性和丰度。基于两组差异表达的肠道菌群,使用支持向量机建立机器学习模型,采用受试者工作特征曲线和曲线下面积(AUC)评估模型对矮身材的预测效能。结果两组门水平和属水平丰度前10位的肠道菌群占比稍有不同,组间差异无统计学意义。矮身材组肠道菌群香农指数、物种丰富度指数低于对照组(P均<0.05)。两组在Bray-Curtis、Jaccard距离差异均无统计学意义。矮身材组纤毛菌科、副拟杆菌属、盲肠坦纳菌科、直肠真杆菌、阿加索杆菌属、二烯菌科、迪氏菌属、普氏栖粪杆菌相对丰度高于对照组,螺杆菌科、梭状芽胞杆菌、LachnospiraceaeNC2004group相对丰度低于对照组(P均<0.05)。在种水平筛选与SDS相关的菌群,通过聚类分析发现,油脂分解拟杆菌、贫乏另枝菌、口腔放线菌、瘤胃球菌之间存在明显正相关。基于种水平差异表达肠道菌群构建支持向量机预测模型,经过参数调整与优化,预测模型1预测矮身材的AUC为0.856(95%CI 0.7267~0.9905);进一步调整其他可能影响患儿身高的常见因素(BMI、遗传身高、是否早产、出生身长、出生后喂养方式等),调整后预测模型2预测矮身材的AUC为0.788(95%CI 0.6361~0.9397)。两种模型中,对预测影响较大的肠道菌群包括组织普雷沃菌、油脂拟杆菌、马赛斯托克菌和弗斯莫尔登乳杆菌。结论矮身材患儿与正常儿童肠道菌群存在差异表达,基于差异表达肠道菌群建立的预测模型有助于早期识别矮身材发生风险。

Objective To analyze the differences in gut microbiome expression between children with short stature and those with normal height,and to explore the relationship between changes in gut microbiome and short stature.Methods Sixty-six children with short stature were selected in the short stature group,and 56 healthy children were included in the control group.The diversity and abundance of gut microbiome were analyzed and compared between the two groups.Based on the differentially expressed gut microbiome between the two groups,a machine learning model was established using support vector machines,and the model's predictive performance for short stature was evaluated using receiver operating characteristic(ROC)curve and area under the curve(AUC).Results The proportions of the top 10 gut microbiome at the phylum and genus levels were slightly different between the two groups,but the differences were not statistically significant.The Shannon index and species richness index of gut microbiome in the short stature group were lower than those in the control group(both P<0.05).There were no statistically significant differences in Bray-Curtis or Jaccard distances between the two groups.The relative abundances of Leptotrichiaceae,Paraprevotellaceae,Tannerellaceae,Eubacteriumrectale,Agathobacter,Dietziaceae,Dietzia,and Faecalibacterium prausnitzii in the short stature group were higher than those in the control group,while the relative abundances of Helicobacte raceae,Clostridium sensu stricto,and Lachnospiraceae NC2004 group were lower(all P<0.05).At the species level,cluster analysis of microbiome related to SDS revealed significant positive correlations among Bacteroides oleiciplenus,Alistipes,Actinomyces oris,and Ruminococcus.Based on the differentially expressed gut microbiome at the species level,an support vector machine prediction model was constructed.After parameter adjustment and optimization,the AUC of prediction model 1 for short stature was 0.856(95%CI:0.7267-0.9905);after further adjusting for other common factors that affected children's height(BMI,genetic height,premature birth,birth length,feeding method after birth,etc.),the AUC of prediction model 2 for short stature was 0.788(95%CI:0.6361-0.9397).In both models,the gut microbiome with greater influence on prediction included Prevotella histicola,Bacteroides oleiciplenus,Marvinbryantia,and Fusicatenibacter saccharivorans.Conclusions There are differences in gut microbiome expression between children with short stature and normal children.The prediction model based on differentially expressed gut microbiome shows good predictive performance for short stature.