A new method for classifying different phenotypes of kidney transplantation

详细信息    查看全文

• 作者：Dong Zhu ; Zexian Liu ; Zhicheng Pan ; Mengjia Qian…
• 刊名：Cell Biology and Toxicology
• 出版年：2016
• 出版时间：August 2016
• 年：2016
• 卷：32
• 期：4
• 页码：323-332
• 全文大小：2,287 KB
• 刊物类别：Biomedical and Life Sciences
• 刊物主题：Life Sciences
  Cell Biology
  Pharmacology and Toxicology
  Biochemistry
  
• 出版者：Springer Netherlands
• ISSN：1573-6822
• 卷排序：32

文摘

For end-stage renal diseases, kidney transplantation is the most efficient treatment. However, the unexpected rejection caused by inflammation usually leads to allograft failure. Thus, a systems-level characterization of inflammation factors can provide potentially diagnostic biomarkers for predicting renal allograft rejection. Serum of kidney transplant patients with different immune status were collected and classified as transplant patients with stable renal function (ST), impaired renal function with negative biopsy pathology (UNST), acute rejection (AR), and chronic rejection (CR). The expression profiles of 40 inflammatory proteins were measured by quantitative protein microarrays and reduced to a lower dimensional space by the partial least squares (PLS) model. The determined principal components (PCs) were then trained by the support vector machines (SVMs) algorithm for classifying different phenotypes of kidney transplantation. There were 30, 16, and 13 inflammation proteins that showed statistically significant differences between CR and ST, CR and AR, and CR and UNST patients. Further analysis revealed a protein-protein interaction (PPI) network among 33 inflammatory proteins and proposed a potential role of intracellular adhesion molecule-1 (ICAM-1) in CR. Based on the network analysis and protein expression information, two PCs were determined as the major contributors and trained by the PLS-SVMs method, with a promising accuracy of 77.5 % for classification of chronic rejection after kidney transplantation. For convenience, we also developed software packages of GPS-CKT (Classification phenotype of Kidney Transplantation Predictor) for classifying phenotypes. By confirming a strong correlation between inflammation and kidney transplantation, our results suggested that the network biomarker but not single factors can potentially classify different phenotypes in kidney transplantation.KeywordsAllograft rejectionInflammationIntracellular adhesion molecule-1Kidney transplantationPLS-SVMs