Background: Acute kidney injury (AKI) is a common clinical syndrome with limited methods of treatment and diagnosis. Although several molecules associated with AKI have been discovered, molecular mechanisms underlying AKI still remain unclear. Weighted gene co-expression network analysis (WGCNA) is a novel method to uncover the relationship between co-expression genes and clinical traits at the system level.Methods: First, by employing WGCNA in transcriptional data on 30 patients with well/poor functioning kidney graft, we identified two co-expression modules that were significantly related to serum creatinine (SCr). Second, based on the modules, potential small molecular compound candidates for developing targeted therapeutics were obtained by connectivity map analysis. Furthermore, multiple validations of expression in space/time were carried out with two classical AKI models in vivo and other five databases of over 152 samples.Results: Two of the 14 modules were found to be closely correlated with SCr. Function enrichment analysis illustrated that one module was enriched in the immune system, while the other was in the metabolic process. Six key renal function-related genes (RFRGs) were finally obtained. Such genes performed well in cisplatin-induced or cecal ligation and puncture-induced AKI mouse models.Conclusion: The analysis suggests that WGCNA is a proper method to connect clinical traits with genome data to find novel targets in AKI. The kidney tissue with worse renal function tended to develop a “high immune but low metabolic activity” expression pattern. Also, ACSM2A, GLYAT, CORO1A, DPEP1, ALDH7A1, and EPHX2 are potential targets of molecular diagnosis and treatment in AKI.