Lauren classification is a pathology-based gastric cancer (GC) subtyping system, which is widely used in the clinical treatment of patients with GC. However, genome-scale molecular characteristics to distinguish between diffuse (DF) and intestinal (IT) GC remain incompletely characterized, particularly at the transcriptional regulatory level. In the present study, gene regulatory networks were constructed using the Passing Attributes between Networks for Data Assimilation (PANDA) algorithm for DF, IT and mixed GC. The results indicated that >85% of transcription factor (TF)-target edges were shared among all three GC subtypes. In TF enrichment analysis, 13 TFs, including nuclear transcription factor Y subunit α (NFYA) and forkhead box L1, were activated in DF GC, whereas 8 TFs, including RELA proto-oncogene and T-cell leukemia homeobox 1 (TLX1), were activated in IT GC. Out of these identified TFs, NFYA [Hazard ratio (HR) (95% confidence interval, CI)=0.560 (0.349, 0.900), P=0.017] and sex determining region Y [HR (95% CI)=0.603 (0.375, 0.969), P=0.037] were identified as independent prognostic factors in DF GC, but not in IT GC, whereas TLX1 [HR (95% CI)=0.547 (0.321, 0.9325), P=0.027] was identified as an independent prognostic factor in IT GC, but not in DF GC. Verification at the cellular level was also performed; interference of NFYA expression using small interfering RNA in MGC803 cells (DF GC-derived cells) markedly inhibited cell growth and colony formation. Similar effects were also detected in SGC-7901 cells (IT GC-derived cells), but to a lesser extent. In conclusion, identified gene regulatory networks differed between distinct GC subtypes, in which the same TFs had different biological effects. Specifically, NFYA was identified as a DF subtype-specific independent prognostic factor in GC.