马 伟,李丹丹,张常建,陈伟浩,熊 鸣,乔媛媛,李 军.基于全转录组的心力衰竭预测标志物[J].转化医学杂志,2021,10(6):349-353
基于全转录组的心力衰竭预测标志物
A global transcriptome-based biomarker for the robust classification of heart failure
  
DOI:
中文关键词:  心力衰竭  标志物  最优密码子  蛋白质氧含量  随机森林
英文关键词:Heart failure  Biomarker  Optimal codon  Protein oxygen content  Random Forest
基金项目:国家自然科学基金(31500756);解放军总医院第六医学中心创新培育基金(CXPY202007)
作者单位
马 伟 解放军总医院第六医学中心 
李丹丹 中山大学孙逸仙纪念医院 
张常建 解放军总医院第六医学中心 
陈伟浩 解放军总医院第三医学中心卫勤部 
熊 鸣 解放军总医院第六医学中心 
乔媛媛 解放军总医院第六医学中心 
李 军 解放军总医院第六医学中心 
摘要点击次数: 72
全文下载次数: 83
中文摘要:
      目的 挖掘基于转录组的预测心力衰竭(heart failure,HF)的标志物。方法 本课题组开发了一个基于全转录组的HF预测标志物,整理转录本对应的蛋白质的氧原子含量,然后计算转录本蛋白质氧原子含量与转录本表达水平的相关性,通过随机森林模型训练预测模型。结果 对于训练集,预测模型5折交叉验证的准确率、阳性预测值和阴性预测值分别为83.6%、85.8%和81.5%。训练集的准确率、阳性预测值和阴性预测值分别为77.8%、80.0%和75.0%。在另外一个单独寻找的测试集,准确率、阳性预测值和阴性预测值分别为86.8%、86.7%和87.5%。结论 本课题组开发的基于全转录组的HF预测标志物具有非常稳定的预测性能,而且准确率高,可以用于HF预测。
英文摘要:
      Objective To explore a Heart failure (HF) prediction biomarker based on transcriptome. Method We presented a global-transcriptome based HF biomarker, which integrated the expression level of all the transcripts and the protein oxygen atom content (OAC) of the corresponding transcripts. We trained a random forest classifier using the correlation coefficients between gene expression level and OAC. Result We presented a global-transcriptome based HF biomarker, which integrated the expression level of all the transcripts and the protein oxygen atom content (OAC) of the corresponding transcripts. We trained a random forest classifier using the correlation coefficients between gene expression level and OAC. The accuracy, positive predictive value (PPV) and negative predictive value (NPV) are 83.6%, 85.8% and 81.5% from the five-fold cross-validation, and 77.8%, 80.0%, and 75.0% on the testing dataset, respectively. Moreover, on a totally independent testing dataset, the classifier achieved an accuracy of the 86.8%, with a PPV of 86.7% and a NPV of 87.5%. Conclusion The global transcriptome-based biomarker we developed show robust prediction performance across different datasets, which further suggests that this biomarker could be useful for the classification of HF.
查看全文  查看/发表评论  下载PDF阅读器
关闭