Identification of MYC and STAT3 for early diagnosis based on the long noncoding RNA-mRNA network and bioinformatics in colorectal cancer

Figures

• 

  Figure 1. Screening method. Diagram “K” means the number of records, by screening the literature, we obtained 360 eligible publications on lncRNA and colorectal cancer in this article.

• 

  Figure 2. ceRNA network. We mapped this network through the relationships regulated by the ceRNA network by summarizing the relationships between the lncRNAs, miRNAs and mRNAs we screened for.

• 

  Figure 3. Molecular functions of lncRNAs in CRC progression. Due to the significant differences in lncRNA as oncogenes or oncogenes suppressors, they can be graphically depicted in a scale. On the left side of the scale (red), lncRNAs act as oncogenes in higher numbers, some of which are also involved in signaling pathways; on the right side of the scale (green), lncRNAs act as tumor suppressors in lower numbers.

• 

  Figure 4. The function of lncRNA. We collected and identified a number of differentially expressed lncRNA functions.

• 

  Figure 5. Hub gene gained from CytoNCA. Hub genes were calculated in CytoNCA using Betweenness Centrality and ranked according to their score scores (Appendix 1), the top 28 scores are shown in this figure. The genes in the middle red are usually considered to be the highest scoring and the larger they are the higher the score proved to be.

• 

  Figure 6. Hub gene gained from MCC. Hub genes were calculated in Cytohubba using MCC, and ranked in order of their relevance (Appendix 2), with the top 30 scores as shown in this figure. These genes in the inner circle are considered to be the genes with higher scores as calculated by MCC, and the outer circle has lower scores in comparison.

• 

  Figure 7. Hub gene gained from MCODE. Hub genes were derived from MCODE calculations (Appendix 3) and ranked based on the size of the scores, with the top 30 scores as shown in this figure. These genes in the inner circle are considered to be the genes with higher scores as calculated by MCODE, and the outer circle has lower scores in comparison.

• 

  Figure 8. Intersection of core genes. Taking the top ten with high MCC scores, the top ten with high BC scores, and the top 29 validated by MCODE to obtain the intersection, which in turn identified our five hub genes as TP53, MYC, STAT3, KRAS, and NOTCH1.

• 

  Figure 9. Differential expression of TP53,MYC,STAT3,NOTCH1 and KRAS in colorectal tissues. (A) shows that TP53 is upregulated in colorectal cancer(p<0.001). (B) shows that STAT3 is low expressed in colorectal cancer(p<0.05). (C) shows NOTCH1 is upregulated in colorectal cancer(p<0.001). (D) shows that MYC has upregulation in colorectal cancer(p<0.001). (E) shows that KRAS has downregulation in colorectal cancer(p<0.001). (*, p<0.05；**, p<0.01; ***, p<0.001).

• 

  Figure 10. GEPIA verification. (A) shows that TP53 is significantly upregulated in colorectal cancer(p<0.05).(B) shows that MYC is significantly upregulated in colorectal cancer(p<0.05). (C) shows STAT3 is low expressed in colorectal cancer(p<0.05). (D) and (E) show that KRAS and NOTCH1 have been upregulated in colorectal cancer. (*, p<0.05).

• 

  Figure 11. qRT-PCR results. The difference in the expression of Hub genes in normal colorectal and colorectal cancer cells. (A) showed that the MYC was upregulated in colorectal cancer (p=0.0018); (B) showed that STAT3 was downregulated in colorectal cancer cells (p=3.3e-05).

• 

  Figure 12. ROC curve results. (A) ROC curve analysis showed the predictive performance of the MYC model. The AUC value was 0.937 (95% CI: 0.874-1.000), indicating that the model has good predictive accuracy. (B) ROC curve analysis showed the predictive performance of the STAT3 model. the AUC value was 0.722 (95% CI: 0.572-0.872), indicating that the model has good predictive accuracy.