Objective — to identify the most reliable expression prognostic markers of lymph node-negative locally advanced prostate cancer (LAPC). Material and methods. A total of 130 samples of primary tumors from the representatives of the Caucasian race (The data of the Cancer Genome Atlas (TCGA) consortium) underwent transcriptome analysis. For differential expression analysis, the samples were divided into favorable (n=109) and unfavorable (n=21) prognosis groups according to the absence or occurrence of biochemical recurrence throughout the follow-up (at least 3 years). Only the genes that had been previously identified to be associated with prostate cancer (PC) progression were considered to improve the reliability of results in the future. Potential prognostic markers were also identified for the most common molecular subtype of PC (41 and 10 patients with favorable and unfavorable prognosis, respectively). Results. There were potential markers for the unfavorable prognosis of lymph node-negative LAPC: three genes (TWIST1, TUBB3 and CHAT) that were more highly expressed in the biochemical relapse group than in the comparison group, as well as the CYP1B1, IGSF1, EDN3, MSMV, and SERPINA3 genes that demonstrated reduced mRNA expression. Analysis of the molecular subtype characterized by the formation of TMPRSS2-ERG chimeric transcript in tumor cells showed that the potential markers of unfavorable prognosis were 6 genes: MAPK8IP2, LRFN2, CHGA, TUBB3, MYT1, and CHAT, which were more highly expressed in the biochemical recurrence group than in the comparison group, as well as the MYBPC1 and PAGE4 genes that demonstrated reduced expression. The increased expression of the CHAT and TUBB genes is a potential marker of unfavorable prognosis for both the combined sample of lymph node-negative LAPC and its most common molecular subtype. Conclusion. The authors identified a number of genes (MARK8IP2, LRFN2, CHGA, MYT1, TUBB3, and CHAT), the expression level of which correlated with relapse-free survival. The obtained results can form the basis for a diagnostic panel that after clinical testing would make it possible to create a Russian test system to predict the course of PC, which could consider the molecular subtype of the tumor.