Abstract
A complex disease, especially cancer, always has pre-deterioration stage during its progression, which is difficult to identify but crucial to drug research and clinical intervention. However, using a few samples to find mechanisms that propel cancer crossing the pre-deterioration stage is still a complex problem. In this study, we successfully developed a novel single-sample model based on node entropy with a priori established protein interaction network. Using this model, critical stages were successfully detected in simulation data and four TCGA datasets, indicating its sensitivity and robustness. Besides, compared with the results of the differential analysis, our results showed that most of dynamic network biomarkers identified by node entropy, such as NKD2 or DAAM1, located in upstream in many important cancer-related signaling pathways regulated intergenic signaling within pathways. We also identified some novel prognostic biomarkers such as PER2, TNFSF4, MMP13 and ENO4 using node entropy rather than expression level. More importantly, we found the switch of non-specific pathways related to DNA damage repairing was the main driven force for cancer progression. In conclusion, we have successfully developed a dynamic node entropy model based on single case data to find out tipping point and possible mechanism for cancer progression. These findings may provide new target genes in therapeutic intervention tactics.
Highlights
Cancer has become the most difficult disease in all medical fields and is one of the leading causes of death in the world (Nordin et al, 2019)
A sudden increases of the single-sample node entropy (SNE) represented the imminent critical point when the system was near the parameter value p = 0 (Figure 2B)
To further clarify the association between these core genes fluctuated in critical stages and the differential expressed gene, we focused on several pathways most related to cancer progression and treatment (Perry et al, 2020; Yu et al, 2020)
Summary
Cancer has become the most difficult disease in all medical fields and is one of the leading causes of death in the world (Nordin et al, 2019). Many targeted drugs have been developed based on this strategy, it must be acknowledged that such result-oriented molecular targets still have failure cases, even the most popular immune checkpoint treatments (Le et al, 2015; Sharma and Allison, 2015). Such studies based on pairwise differences can only help us to recognize the results that have occurred. A recent study suggests that most oncogenetic pathways which support cancer cell survival are already over-active in seeming normal stages (Zielinska and Katanaev, 2019). We think mathematics and informatics strategy may help us solve these problems
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
