Identification of NPPB co-expression genes
A total of 577 negatively co-expressed genes and 457 positively co-expressed genes in the DHF group were identified, along with 666 negatively co-expressed genes and 422 positively co-expressed in the nDHF group. Figure 2 portrays 106 negatively and 63 positively co-expressed genes in both patient types, whereby 173 intersection co-expression genes were screened out. Interestingly, of these intersection co-expression genes, we found 3 genes (CENPBD1P1, KHDRBS3, and PHOX2B) that were positively co-expressed with NPPB in patients with T2DM, but negatively co-expressed in patients without T2DM, and 1 gene (NQO1) that was negatively co-expressed with NPPB in patients with T2DM, but positively co-expressed in patients without T2DM.
Functional GO and KEGG pathway enrichment analyses
GO analyses revealed 41 BPs, 20 CCs, and 13 MFs in the DHF group, and 61 BPs, 16 CCs, and 13 MFs in the nDHF group (details in Tables S1-S2). Due to the excessive number of enrichment analyses, the top seven BPs, CCs, and MFs were selected for visualization with P < 0.05 (Figure 3A and Figure 3C). Further, there were 10 BPs (fatty acid beta-oxidation, oxidation-reduction process, metabolic process, mitochondrial respiratory chain complex I assembly, glyoxylate metabolic process, ubiquinone biosynthetic process, positive regulation of cell growth, tricarboxylic acid cycle, cell adhesion, and aerobic respiration), 8 CCs (mitochondrial inner membrane, extracellular space, mitochondrion, extracellular matrix, myelin sheath, extracellular exosome, Z disc, and mitochondrial matrix), and 3 MFs (growth factor activity, protein binding, and electron carrier activity) enriched in both patient groups. There were 41 identified pathways in patients with T2DM (Figure 3B) and 22 in patients without T2DM (Figure 3D) (details in Table S3-S4). Moreover, common pathways are shown in Table 1.
The analyses further identified 16 BPs, 8 CCs, and 3 MFs that were enriched by intersectional co-expression genes in both patient groups (Figure 3E), and these genes mainly clustered in the following 8 pathways: the citrate cycle (TCA cycle), carbon metabolism, biosynthesis of antibiotics, malaria, glyoxylate metabolism, dicarboxylate metabolism, cardiac muscle contraction, and African trypanosomiasis (Figure 3F) (details in Table S5).
PPI network construction and hub gene identification
As Figure 4 shows, the interactions among intersectional co-expression genes were displayed by a PPI network with 273 edges and 170 nodes. This finding was saved in TSV format and then imported into Cytoscape for visualization. With a cutoff criterion of a degree that is > 5 and a K-core > 5, only one module with 4 BPs (the tricarboxylic acid metabolic process, citrate metabolic process, tricarboxylic acid cycle, and aerobic respiration) and 3 pathways (citrate cycle, malaria parasite metabolic pathway, and AGE-RAGE signaling pathway in diabetic complications) significantly enriched in was identified. With the degree ratio ranking method, the top 10 hub genes of this PPI network were also identified (CS, DECR1, ACO2, BGN, TIMP1, CTGF, VCAN, SERPINE1, SDHC, and CCL2). With the same cutoff criterion, a PPI network that consists of 953 nodes and 4,946 edges of NPPB co-expression genes in the DHF group, and a PPI network of 1,009 nodes and 4,245 edges in the nDHF group were also constructed. The top 10 hub genes of the former were CYCS, FN1, CS, DECR1, ACO2, ATP5A1, NDUFAB1, EGF, ATP5H, and ATP5C1, while the later were CS, DECR1, BGN, TIMP1, ACO2, CTGF, VCAN, SERPINE1, CCL2, and SDHC (Figures 5 and 6, respectively). Among these three hub genes clusters, ACO2, CS, and DECR1 were the most common.