Machine learning with in silico analysis markedly improves survival prediction modeling in colon cancer patients

Abstract

Background Predicting the survival of cancer patients provides prognostic information and therapeutic guidance. However, improved prediction models are needed for use in diagnosis and treatment. Objective This study aimed to identify genomic prognostic biomarkers related to colon cancer (CC) based on computational data and to develop survival prediction models. Methods We performed machine‐learning (ML) analysis to screen pathogenic survival‐related driver genes related to patient prognosis by integrating copy number variation and gene expression data. Moreover, in silico system analysis was performed to clinically assess data from ML analysis, and we identified RABGAP1L , MYH9 , and DRD4 as candidate genes. These three genes and tumor stages were used to generate survival prediction models. Moreover, the genes were validated by experimental and clinical analyses, and the theranostic application of the survival prediction models was assessed. Results RABGAP1L , MYH9 , and DRD4 were identified as survival‐related candidate genes by ML and in silico system analysis. The survival prediction model using the expression of the three genes showed higher predictive performance when applied to predict the prognosis of CC patients. A series of functional analyses revealed that each knockdown of three genes reduced the protumor activity of CC cells. In particular, validation with an independent cohort of CC patients confirmed that the coexpression of MYH9 and DRD4 gene expression reflected poorer clinical outcomes in terms of overall survival and disease‐free survival. Conclusions Our survival prediction approach will contribute to providing information on patients and developing a therapeutic strategy for CC patients.

Survival Prediction Landscape: An In-Depth Systematic Literature Review on Activities, Methods, Tools, Diseases, and Databases

“…A deeper analysis of existing survival predictors reveals that among the 74 studies 54 utilized publicly accessible data from three key databases: the Cancer Genome Atlas Program (TCGA) 17 , NCI Genomic Data Commons (GDC) 18 , and the Gene Expression Omnibus (GEO) 31, 32, 72, 73, 80, 82, 87, 90, 91, 130, 131 . Apart from public databases, there also exist private databases that have been utilized in existing survival prediction studies 66,75,81,112,113,117,118 . However, these private databases often restrict data access and may require extensive research proposals for data retrieval.…”

“…Additionally, Network based methods include network based stratification (NBS) 83 , weighted correlation network analysis (WGCNA) 86 , canonical correlation analyses (CCA) 67 , patient similarity networks 38 , and neighborhood component analysis (NCA) 23 . Dimensionality reduction methods include non-negative matrix factorization (NMF) 40 , autoencoders (AEs) 28 , variational autoencoders (VAEs) 43 , principal component analysis (PCA) 39 , and dominant effect of the cancer driver genes (DEOD) 75,132 . Moreover, clustering methods comprise Kruskal-Wallis and Gaussian clustering 131 , hierarchical clustering 82 , and Guassian clustering 131 .…”

“…Table 8 provides information about 44 diseases and the corresponding survival prediction algorithms utilized in these diseases. A deeper analysis of Table 8 shows that Cox-PH and lasso Cox-PH models have been extensively utilized for disease specific survival prediction i.e., ASCVD 29,111 , bladder cancer 40,82 , colorectal cancer [74][75][76][77] , hepatocellular carcinoma 43,86,87 , ovarian cancer [88][89][90]103 , lung adenocarcinoma 101 , heart failure 118 , HER2-negative metastatic breast cancer 67 , pancreatic cancer 26,71 , trauma 120 , nasopharyngeal carcinoma 66 , triple-negative breast cancer 68 , lymphoma 85 , breast cancer 40,81,82 , ovarian cancer [88][89][90]103 , and lower-grade glioma 80 , cardiovascular disease 112,[114][115][116][117] , invasive ductal carcinoma 70 , liver transplantation 119 , gastric cancer 42 , lung cancer 27 , esophageal squamous cell carcinoma 79 , glioma 69 , and liver cancer …”

“…DL model DeepSurv, has been utilized in 5 studies related to gastrointestinal cancer 30 , ASCVD 111 , NSCLC 97 . On the other hand, in the analyzed survival predictive pipelines less frequently utilized methods are i.e., survival SVM 79,95,120 , partial logistic regression 70,75 , log hazard net 75,104 , boosting 41,112 , stepCox 86 , elastic net 95 , CNNcox 104 , DeepOmix 104 , ordinal Cox-PH 78 , DeepHit 112 , and linear multitask logistic regression (MTLR) 112 .…”

“…The choice of omics type hinges on the specific disease under investigation, as indicated by the disease-wise distribution of omics types in Figure 5. Out of 9 omics types, mRNA, CNV, miRNA, and methylation have been the most commonly utilized modalities for 33 cancer subtypes i.e., breast cancer 14,23,68,74,90,[98][99][100] , pancancer 24,91,[105][106][107][108]131 , colon can- cer 39,[74][75][76][77] , lung adenocarcinoma 27,101,102 , and ovarian cancer 72,84,[88][89][90]103 . In addition, mutation data has been utilized for 7 cancer subtypes namely, adult diffuse glioma 69 , breast cancer 23 , cervical cancer 73 , non-small cell lung cancer 95 , ovarian cancer 103 , and pancreatic cancer 32 .…”