The details of our selection process are shown in Supplementary Fig. 1. LN was completely distinguished from HC by PCA analysis (Fig. 1a). An aggregate of 201 DEGs were acquired, of which 31 DEGs were upregulated and 170 was downregulated (Fig. 1b). As the adjusted p-value decreased, DEGs increased in ranking in this experiment. As shown in Fig. 1c, the R heatmap software was used to draw a heatmap of the 31 upregulated and 50 downregulated DEGs.
GO term and KEGG pathway enrichment analysis of DEGs
With an adjusted p-value of < 0.05, the GO term and KEGG pathway enrichment analyses of upregulated and downregulated genes were acquired, respectively. The results of the GO term in LN are shown in Fig. 2a–d. The visual analysis results of the KEGG enrichment of DEGs in LN are shown in Fig. 3a and b. GO analyses revealed that DEGs were enriched in biological processes (Fig. 2b),such as chemical response, cellular response to chemical stimulus, and response to external stimulus; molecular functions (Fig. 2c), such as transporter activity, transition metal ion binding, and anion transmembrane transporter activity; cellular components (Fig. 2d), such as extracellular region, vesicle, and extracellular region. These DEGs were enriched in KEGG pathways, including the peroxisome proliferator-activated receptor (PPAR) signaling pathway, phenylalanine metabolism, and renin-angiotensin system (Fig. 2e and f). The pathway map for targeted LN (Supplementary Fig. 2) was described using KEGG pathway enrichment. The significantly enriched terms and pathways may help us to further explore the role of DEGs in LN.
Analyzing DEGs in LN using a PPI network
To construct PPI networks, DEG expression products in LN were constructed using the STRING database. After deleting all isolated and partially disconnected nodes, as shown in Fig. 3a, an integrated network was built. The 30 most significant genes are shown in Fig. 3b.
Identification and Validation of Hub Genes
Finally, Venn diagram was adopted to illustrate the intersection of five algorithms (Fig. 3c, Supplementary Table 1). Based on the Immunology Database and Analysis Portal (ImmPort), two hub genes, FOS, and insulin-like growth factor 1(IGF1), were screened out. The number of nodes in the FOS network was 32, and the number of nodes in the IGF1 network was 32, both of which were > 30, indicating that the two genes were more strongly associated with other genes in the PPI network.
Association between the Hub Genes and Clinical Features of LN
Using Nephroseq v5, the expression of the hub genes (FOS and IGF1) showed the difference between LN and HC; both hub genes were downregulated in the LN renal tissues, compared with the HC renal tissues (Fig. 4a and b). In addition, the expression of FOS in LN renal tissue samples was negatively correlated with GFR (Fig. 4c and d), which is one of the important indices for measuring renal function.
Diagnostic Effectiveness of Biomarkers for LN
The normalized dataset was used to verify the diagnostic specificity and sensitivity of the biomarkers for LN by ROC analysis. As shown in Fig. 4e and f, the AUC values of FOS and IGF1 were 0.786 (95% CI, 0.676–0.878) and 0.885 (95% CI, 0.802–0.951), respectively. Therefore, FOS and IGF1 could diagnose LN.
Immune infiltration analyses
In studying immune cell infiltration, 155 tissues, including 25 HC and 130 LN tissues, were identified (Supplementary Fig. 3). PCA (Fig. 5a) showed that a significant difference (P < 0.05) in immune cell infiltration existed between LN with HC. The results of the immune cell infiltration analysis indicated that eosinophils were not ex-pressed in all tissues. As shown in Fig. 5b, 10 types of differentially infiltrating immune cells between the LN and HC groups were present. A clear correlation existed between increasing B cell memory and T cell gamma delta in LN, which suggests that humoral immunity plays an essential role in the mechanism of LN. Thereafter, the percentage of the 22 types of infiltrating immune cells was analyzed (Fig. 5c). Every differentially infiltrating immune cell was assessed using the Wilcoxon rank test. Activated and resting DCs were lower in LN than M1 macrophages and activated NK cells. This finding could reflect the heterogeneously inflammatory character of LN (Fig. 5d).
Correlation Analysis Between Key Biomarkers and Infiltration-Related Immune Cells
Based on the results of the correlation analysis, FOS showed a positive correlation with activated mast cells (R = 0.27, P = 8e-04) and showed a negative correlation with resting mast cells (R = − 0.23, P = 0.0037). IGF1 had a positive correlation with activated DCs (R = 0.33, P = 3.4e-05) and a negative correlation with monocytes (R = − 0.22, P = 0.0053) (Supplementary Fig. 4a–e).
Identified and analysis of miRNAs in LN
According to the HMDD database, it was reported that 34 miRNAs were associated with LN. Based on Tarbase, we constructed the hub genes-miRNAs network and the LN-related miRNAs were shown in red on the network (Fig. 6). Among the miRNAs targeted for the hub genes, hsa-miR-155-5p, hsa-miR-1-3p, hsa-miR-182-5p and has-miR-16-5p were the top miRNAs that considered targeted for two hub genes. We crossed the miRNA screened from HMDD with these three RNAs and found the hub miRNA, has-mir-155-5p.
Landscape of m 6 A RNA methylation regulators in LN.
Wilcoxon rank-sum test (for unpaired comparisons) was used to compare the m6A regulators’ expression level between LN and normal cases. Unpaired comparisons showed that CBLL1, METTL5, WTAP, RBM15/15B, ZC3H13, ZCCHC4, FTO, YTHDF1/3, YTHDC1/2, and IGFBP1/2 were all downregulated in LN (Fig. 7A). Based on the correlation analysis between m6A genes and hub genes, there was consistent variation in IGFBP3, YTHDF3, ZCCHC4, RBM15 and FTO in IGF1 and FOS (Figs. 7B-C). To further elucidate the roles of these m6A markers in LN, correlation analysis between markers and GFR was carried out in the Nephroseq database (Figs. 8). RBM15, RBM15B, WTAP, YTHDF1 and ZC3H13 were all negatively correlated with GFR, indicating that these genes may aggravate kidney damage in patients with LN.