RESUMEN
Objective:To explore the poor early liver graft function(PEGF)-related biomarkers and establish a genomic model for PEGF prediction specific to liver transplantation(LT)with allografts of donation after brain death(DBD).Methods:By data-mining a public GSE23649 dataset from the database of Gene Expression Omnibus(GEO), key PEGF-related genes in DBD liver biopsies after 2h reperfusion were identified by differential expression analysis.And LASSO-penalized Logistic regression model was utilized for selecting an optimal gene set.Receiver operating characteristic curves with its area under the curve(AUC)and a nomogram were generated for evaluating and visualizing its predictive capability for PEGF.Gene set enrichment analysis(GSEA) was performed for exploring the biological pathways underlying PEGF.Results:Six key PEGF-related genes in DBD-LT were initially identified, including 4 up-regulated genes(HBB, PFDN5, RPS3A & RPS5)and 2 down-regulated genes(RPL22 & FAM62B). A six-mRNA-based risk-scoring model was further established with an excellent predictive capability(AUC=1.000, P=0.0008). Four PEGF-related biological pathways in DBD livers, such as "VEGF" and "natural killer cell-mediated cytotoxicity" , were identified by GSEA(all P<0.05). Conclusions:The genomic model may effectively predict the likelihood of PEGF immediately after DBD-LT or even prior to transplantation in the context of normothermic machine perfusion.