Limits...
Competing risks data analysis with high-dimensional covariates: an application in bladder cancer.

Tapak L, Saidijam M, Sadeghifar M, Poorolajal J, Mahjub H - Genomics Proteomics Bioinformatics (2015)

Bottom Line: Analysis of microarray data is associated with the methodological problems of high dimension and small sample size.By fitting to the Fine and Gray model, eight genes were highly significant (P<0.001).Among them, expression of RTN4, SON, IGF1R, SNRPE, PTGR1, PLEK, and ETFDH was associated with a decrease in survival time, whereas SMARCAD1 expression was associated with an increase in survival time.

View Article: PubMed Central - PubMed

Affiliation: Department of Biostatistics and Epidemiology, School of Public Health, Hamadan University of Medical Sciences, Hamadan 65175-4171, Iran.

No MeSH data available.


Related in: MedlinePlus

The prediction error curves for bladder cancer dataClinical model used age, sex, stage, grade and treatment as predictors. The elastic net, Lasso, and boosting used microarray features in addition to the clinical parameters as predictors.
© Copyright Policy - CC BY-NC-ND
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4563215&req=5

f0010: The prediction error curves for bladder cancer dataClinical model used age, sex, stage, grade and treatment as predictors. The elastic net, Lasso, and boosting used microarray features in addition to the clinical parameters as predictors.

Mentions: To assess predictive performance, the median area under the curve (AUC) was calculated and plotted for each method. The results are presented in Figure 1. The average median AUC (across all time points) were 0.808, 0.695, and 0.729 for the elastic net, Lasso, and boosting methods, respectively. As shown in Figure 1, in terms of prediction, the predictive performance of elastic net was superior to the Lasso and boosting in the analysis of this dataset, whereas the predictive performance of boosting was slightly better than the Lasso method. Moreover, bootstrap .632+ prediction error curves were plotted for the three methods (Figure 2). The data clearly indicated that the elastic net outperformed the Lasso and boosting methods, which agreed well with the AUC analysis.


Competing risks data analysis with high-dimensional covariates: an application in bladder cancer.

Tapak L, Saidijam M, Sadeghifar M, Poorolajal J, Mahjub H - Genomics Proteomics Bioinformatics (2015)

The prediction error curves for bladder cancer dataClinical model used age, sex, stage, grade and treatment as predictors. The elastic net, Lasso, and boosting used microarray features in addition to the clinical parameters as predictors.
© Copyright Policy - CC BY-NC-ND
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4563215&req=5

f0010: The prediction error curves for bladder cancer dataClinical model used age, sex, stage, grade and treatment as predictors. The elastic net, Lasso, and boosting used microarray features in addition to the clinical parameters as predictors.
Mentions: To assess predictive performance, the median area under the curve (AUC) was calculated and plotted for each method. The results are presented in Figure 1. The average median AUC (across all time points) were 0.808, 0.695, and 0.729 for the elastic net, Lasso, and boosting methods, respectively. As shown in Figure 1, in terms of prediction, the predictive performance of elastic net was superior to the Lasso and boosting in the analysis of this dataset, whereas the predictive performance of boosting was slightly better than the Lasso method. Moreover, bootstrap .632+ prediction error curves were plotted for the three methods (Figure 2). The data clearly indicated that the elastic net outperformed the Lasso and boosting methods, which agreed well with the AUC analysis.

Bottom Line: Analysis of microarray data is associated with the methodological problems of high dimension and small sample size.By fitting to the Fine and Gray model, eight genes were highly significant (P<0.001).Among them, expression of RTN4, SON, IGF1R, SNRPE, PTGR1, PLEK, and ETFDH was associated with a decrease in survival time, whereas SMARCAD1 expression was associated with an increase in survival time.

View Article: PubMed Central - PubMed

Affiliation: Department of Biostatistics and Epidemiology, School of Public Health, Hamadan University of Medical Sciences, Hamadan 65175-4171, Iran.

No MeSH data available.


Related in: MedlinePlus