2.1 Acquisition of OCI-lncRNAs
13,349 OC lncRNAs as well as 19,262 OC mRNAs were obtained from the TCGA database and 2483 immune genes were obtained from the Immport database. The expression matrix of 7769 OC I-lncRNAs was obtained using Perl scripts and R language. 88 I-lncRNA transcriptome matrices in normal ovarian tissue from GTEX database were screened. The normal ovarian tissue I-lncRNA and OCI-lncRNA expression matrices were combined for differential analysis,540 differentially expressed OCI-lncRNA expression matrices were obtained.
2.2 Identification of immune genes associated with OC prognosis
By co-expression analysis, Pearson correlation test and one-way Cox analysis, we identified 49 PI-lncRNAs that were significantly associated with survival time and survival status (p<0.05). High-Risk PI-lncRNAs(HRPI-lncRNAs), and if the HR was less than 1, the PI-lncRNAs were defined as low-risk PI-lncRNAs (Low-risk PI-lncRNAs, LRPI-lncRNAs). The results were presented in a forest plot (Figure 1).
2.3 OCPI-lncRNA-based prognostic model for OC
LASSO regression analysis and cross-validation optimization were performed on 49 OCPI-lncRNAs, and 27 of these OCCPI-lncRNAs (COLCA1, MINCR, AC068792.1, AL391807.1, AC027020.2, MIRLET7BHG, DLGAP1-AS1, DICER1-AS1, AJ011932.1, AC091806.1, FAM27E3, ALDH1L1-AS2, AC008522.1, AC112491.1, AC134312.1, AC010733.1, FRMD6-AS2, DLGAP1-AS2, PSMB8-AS1, AC012645.4, SLX1A-SULT1A3, AC027348.1, FAM157C, AL121845.4, CHRM3-AS2, PKP4-AS1, U62631.1) were included in the OC prognostic model. Finally,we obtained the risk score formula for the OC prognostic model and the risk score and grouping information (high or low risk-group) for each OC patient, the results of which are presented in the supplementary files genesCoef.txt and riskScore.txt.
2.4 Survival analysis and risk analysis between different risk groups
By comparing the difference in survival prognosis between the high- and low-risk groups by plotting survival analysis curves, the survival rate of the high-risk group was significantly lower than that of the low-risk group (p < 0.001), indicating that our prognostic model could accurately predict the survival prognosis of patients in the high- and low-risk groups based on the differential expression of the screened OCCPI-genes ( Figure 2). By analysing the prognostic risk of patients in the high and low-risk groups, we obtained three sets of risk maps. Patients in the high-risk group had significantly higher risk scores than those in the low-risk group (Figure 3). As the risk score increased, the patients' survival time interval decreased and the number of deaths increased (Figure 4). The expression of OCCPI-lncRNAs in the high-risk group was higher than the expression of CPI-lncRNAs in the low-risk group (Figure 5).
2.5 Validation of the OC prognostic model
By independent prognostic analysis, we used R scripts to draw forest plots for both one-way and multi-way Cox regression analyses. P < 0.001 and HR > 1 for risk scores in both univariate and multifactorial analyses, indicating that risk scores in prognostic models can be used as reliable clinically independent prognostic factors (see Figures 6A,6B). Subsequently, we plotted ROC curves under 1-3 year survival of OC patients to validate the accuracy of the OC prognostic model, and the areas under the ROC curves were 0.767, 0.777 and 0.761 (AUC > 0.70), respectively, indicating that our prognostic model predicted patient survival with high accuracy (Figure 7).
2.6 OCPI-lncRNA immune correlation analysis
The TCGA immune cell infiltration file was downloaded from the TIMER database and combined with the risk data for analysis to obtain a heat map of immune cell correlations (Figure 8A), further refining the analysis of differences in immune function to clarify differences in immune-related functions between the high and low- risk groups (Figure 8B), and yielding high and low-risk groups in MHC-class1, TypeI, and TypeII interferon response All three functions were active (p<0.05). Analysis of immune checkpoint differences revealed differences in the expression of IDO1 and BTLA, two immune checkpoint-related genes, between the high and low-risk groups (P<0.05), and both genes showed upregulation in the low-risk group (Figure 8C).
2.7 Identification of potential target drugs for OC
After risk differential analysis of the transcriptome-wide gene matrix in the high-low risk group in the prognostic model, we calculated and identified four potential OC-targeting drugs (AP24534 (Ponatinib), AUY922 (Luminespib), and Axitinib) by R-script screening with an enrichment score ( Enrichment < 0 and P < 0.01), suggesting that these potential drugs may have inhibitory effects on the expression of PI-lncRNAs. Finally, the 2D and 3D chemical structure formulae of the targeted molecular drugs were visualised through the database PubChem(Figure 9).
A.AP24534 (Ponatinib)