Skin cutaneous melanoma (SKCM) is a cancer with serious global impact. Long non-coding RNA was previously found to be associated with tumor prognosis. This research focuses on long intergenic non-protein coding (LINC) RNAs, and correlated protein-coding genes (PCGs), to explore their diagnostic and prognostic value, function and mechanism. Gene expression data was obtained from TCGA and Oncomine for analysis
in total there were 458 cases included in this study. LIN00518 and the 10 most highly correlated PCGs were selected to determine the diagnostic and prognostic value. We undertook bioinformatic analysis with LINC00518 and the prognostic-related PCGs in order to explore their molecular mechanism. The Connectivity Map was carried out for pharmacological target prediction and drug selection. Among the top 10 correlated PCGs, trafficking kinesin protein 2 (TRAK2), epilepsy of progressive myoclonus type 2 gene A (EPM2A) and melanocyte inducing transcription factor (MITF) had significant diagnostic value (all AUC >
0.7, P <
0.05). LINC00518, ras association domain family member 3 (RASSF3), cdk5 and Abl enzyme substrate 1 (CABLES1), kazrin, periplakin interacting protein (KAZN), EF-hand calcium binding domain 5 (EFCAB5) and MITF were significantly associated with prognosis (all adjusted P <
0.05). LINC00518 was associated with cell cycle process, melanogenesis, MAPK signaling pathway, cell division and DNA repair(all P <
0.05). Pharmacological targets analysis suggested results acquired eight potential target drugs. Up-regulation of LINC00518 is significantly associated with poor prognosis. TRAK2, EPM2A and MITF had diagnostic significance. RASSF3, CABLES1, KAZN, EFCAB5 and MITF had prognostic significance. This study provided novel biomarkers for SKCM.