Identification and external validation of the hub genes associated with cardiorenal syndrome through time-series and network analyses

Abstract

Cardiorenal syndrome (CRS), defined as acute or chronic damage to the heart or kidney triggering impairmentof another organ, has a poor prognosis. However, the molecular mechanisms underlying CRS remain largelyunknown. The RNA-sequencing data of the left ventricle tissue isolated from the sham-operated and CRS modelrats at different time points were downloaded from the Gene Expression Omnibus (GEO) database. Genomicdifferences, protein–protein interaction networks, and short time-series analyses, revealed fibronectin 1 (FN1)and periostin (POSTN) as hub genes associated with CRS progression. The transcriptome sequencing data ofhumans obtained from the GEO revealed that FN1 and POSTN were both significantly associated with manydifferent heart and kidney diseases. Peripheral blood samples from 20 control and 20 CRS patients werecollected from the local hospital, and the gene expression levels of FN1 and POSTN were detected by real-timequantitative polymerase chain reaction. FN1 (area under the curve [AUC] = 0.807) and POSTN (AUC = 0.767)could distinguish CRS in the local cohort with high efficacy and were positively correlated with renal and heartdamage markers, such as left ventricular ejection fraction. To improve the diagnostic ability, diagnosis modelscomprising FN1 and POSTN were constructed by logistic regression (F-Score = 0.718), classification tree (F-Score= 0.812), and random forest (F-Score = 1.000). Overall, the transcriptome data of CRS rat models weresystematically analyzed, revealing that FN1 and POSTN were hub genes, which were validated in differentpublic datasets and the local cohort.