1.Identification of DEGs
GSE3365 was selected and underwent DEGs analysis using “Limma” package in R 3.6.3 software.Three ninety-eight DEGs were identified either up- or downregulated in all,including 59 up- and 39 downregulated genes (∣log2FC∣>1 and adjust P-value<0.05). As shown in Figure 2A, all 98 DEGs were plotted that blue ones represented downregulation, red ones indicated upregulation, and gray ones were the rest of the DEGs. Furthermore, the expression levels of all the DEGs were presented in the heatmap (Fig. 2B), and these genes were well clustered between UC and control group.
2. Correlation analysis
The top 10 up- or downregulated DEGs were selected for correlation analysis(∣log2FC∣>1 and adjust P-value<0.05). Figure 2C shows a bitmap of the correlation analysis between the DEGs. Red and green denoted positive and negative correlation, respectively. And the darker the color, the higher the correlation coefficient. According to the classification for Spearman's correlation coefficient (ρ), the absolute value of 0 to 0.10, 0.10 to 0.39, 0.40 to 0.69, 0.70-0.89,and 0.90 to 1.00 represented “negligible” correlation, “weak” correlation, “moderate” correlation, “strong”correlation,and “very strong” correlation, respectively. Furthermore, “ρ=0” indicated “no correlation ” and “ρ=1.00” represented “perfect correlation.”[13]As listed in Figure 2c, ZNF91 had a notable positive correlation with SETD5 (ρ=1.00). However, ZNF91 had a highly negative correlation with AQP9 (ρ=-0.40).ALAS2 had positive correlation with EPB42 (ρ=0.94) as well. Concurrently,ZNF91 and SETD5 also had relatively obvious positive correlation with CELF2 (ρ=0.85 and ρ=0.85).In addition, a stronger negative correlation existed between AQP9 and FCF1 (ρ=-0.53).
3. Functional enrichment analysis of DEGs
To obtain a deeper insight into the biological functions of DEGs,GO annotation and KEGG pathway enrichment analyses were performed. The top 10 enriched GO terms were shown in
Figures 3 and 4. The GO terms were comprised of 3 parts: cellular component (CC), biological process (BP), and molecular function (MF).[14]DEGs of BP were involved in leukocyte migration, acute inflammatory response, neutrophil migration, granulocyte migration, and neutrophil degranulation.CC analysis revealed that DEGs were markedly enriched ecretory granule membrane, external side of plasma membrane , anchored component of membrane ,blood microparticle,and tertiary granule.For MF analysis, the significantly enriched terms were chemokine activity and chemokine receptor binding.Besides, the enriched KEGG pathways as presented in Figure 5, including Cytokine-cytokine receptor interaction,Viral protein interaction with cytokine and cytokine receptor,Pantothenate and CoA biosynthesis,IL-17 signaling pathway and Chemokine signaling pathway.
4. PPI network construction and module analysis
The STRING database was used to identify the PPI pairs. As indicated in Figure 6A, 51 nodes (DEGs) and 218 edges (interactions) were established in the constructed PPI network.
Depending on the degree value, the top 10 hub DEGs were determined. The results shown that CXCL1 and CCL2 were the most crucial genes with the highest degree=26, followed by CXCR1 and FCGR3B at degree=24, C-X-C chemokine receptor type 2 (CXCR2) at degree=22, Prostaglandin G/H synthase 2 (PTGS2) at degree=20, Triggering receptor expressed on myeloid cells 1 (TREM1) at degree=18, Interleukin-1 receptor type 1 (IL1R1) at degree=18, fMet-Leu-Phe receptor (FPR1) at degree=16, and Band 3 anion transport protein (SLC4A1) at degree=14 (Table 1).
Table 1
The information of top 10 hub genes based on their degree value.
Gene name
|
Protein name
|
Expression level
|
Degree
|
Enriched significant modules
|
CXCL1
|
Growth-regulated alpha protein
|
UP
|
26
|
module 3
|
CCL2
|
C-C motif chemokine 2
|
UP
|
26
|
module 2
|
CXCR1
|
C-X-C chemokine receptor type 1
|
UP
|
24
|
module 3
|
FCGR3B
|
Low affinity immunoglobulin gamma Fc region receptor III-B
|
UP
|
24
|
None
|
CXCR2
|
C-X-C chemokine receptor type 2
|
UP
|
22
|
module 2
|
PTGS2
|
Prostaglandin G/H synthase 2
|
UP
|
20
|
module 3
|
TREM1
|
Triggering receptor expressed on myeloid cells 1
|
UP
|
18
|
None
|
IL1R1
|
Interleukin-1 receptor type 1
|
UP
|
18
|
None
|
FPR1
|
fMet-Leu-Phe receptor
|
UP
|
16
|
module 2
|
SLC4A1
|
Band 3 anion transport protein
|
UP
|
14
|
module 1
|
Furthermore, the 3 significant modules (score>4.0) were extracted from the PPI network. Module 1 contained 6 gene nodes, including SELENBP1,SNCA, ALAS2, CA1,EPB42 and SLC4A1 with 28 edges (Fig. 6B). Module 2 contained 4 genes nodes and 12 edges, including Interleukin-1 receptor type 2 (IL1R2), C-C motif chemokine 2 (CCL2), Interleukin-1 receptor type 1 (IL1R1) and C-X-C chemokine receptor type 2 (CXCR2) (Fig. 6C). Module 3 contained 7 genes nodes and 24 edges, including PTGS2, Vascular non-inflammatory molecule 2 (VNN2), Low affinity immunoglobulin gamma Fc region receptor III-B (FCGR3B), C-C motif chemokine 7 (CCL7),CXCL1,CXCR1 and Aquaporin-9 (AQP9) (Fig. 6D). Notably, only one hub DEGs of SLC4A1 was found in module 1. Three hub DEGs of CCL2, CXCR2 and FPR1 were enriched in module 2. Three hub DEGs of CXCL1, CXCR1 and PTGS2 were enriched in module 3. However, FCGR3B ,IL1R1 and TREM1 were not shown in significant modules.
5. Expression level analysis of hub genes
The interaction network between the 10 hub DEGs was constructed by Cytoscape version 3.8.2 plug-in “cytoHubba”based on their degree (Fig. 7A). As demonstrated in Figure 7B, the hub genes expression levels of FPR1,CXCR2,CXCL1,CCL2,CXCR1,FCGR3B,IL1R1,SLC4A1,TREM1,and PTGS2 were markedly upregulated in UC peripheral blood compared to those in non-UC peripheral blood. Furthermore, correlation analysis between the expression levels of the 10 hub genes (FPR1,CXCR2,CXCL1,CCL2,CXCR1,FCGR3B,IL1R1,SLC4A1,TREM1,and PTGS2) was performed by adopting Spearman's correlation analysis. The results demonstrated that there was a positive correlation between the 10 hub genes expression (Fig. 7C). Obviously, CXCR2 had a noteworthy positive correlation with FCGR3B (ρ=0.87).