hdnom app

Nomograms for high-dimensional data, built with ease.


“Hiding within those mounds of data is knowledge that could change the life of a patient, or change the world.” Atul Butte, Professor, Stanford School of Medicine

Please read our data privacy policy before uploading any data.

Read a detailed explanation about the upload data format . An example dataset is provided below.

Sample data x (Comma-separated)      Sample data x (Tab-separated)      Sample data x (Semicolon-separated)
Sample data y (Comma-separated)      Sample data y (Tab-separated)      Sample data y (Semicolon-separated)

Time-Dependent AUC Summary at Evaluation Time Points

Sample data x (Comma-separated)
Sample data x (Tab-separated)
Sample data x (Semicolon-separated)

Sample data y (Comma-separated)
Sample data y (Tab-separated)
Sample data y (Semicolon-separated)

Time-Dependent AUC Summary at Evaluation Time Points

Sample data x (Comma-separated)
Sample data x (Tab-separated)
Sample data x (Semicolon-separated)

Sample data y (Comma-separated)
Sample data y (Tab-separated)
Sample data y (Semicolon-separated)

Please make a nomogram first, then answer the following questions:

Report & Model Download

Please choose one or more from the three types of report according to the analysis you have done. The R model object is also available.

Note: please make sure the parameter settings in the previous steps are what you wanted before generating the report. Feel free to adjust them and regenerate the results if needed.

Basic Report

Nomogram, internal validation, and calibration results

Generate & Download Report

External Validation Report

External validation and calibration results

Generate & Download Report

Model Comparison Report

Model comparison results

Generate & Download Report

R Model Object

Load the model object in R with load("hdnom-model.Rdata") for prediction. Or try hdnom appmaker to make your own nomogram app.


Download R Model Object

Getting started with the hdnom app

This web application helps you build penalized Cox models for high-dimensional data with survival outcomes. All the 9 types of model included in the hdnom package are supported. It streamlined the process of nomogram building, model validation, calibration, comparison, and reproducible report generation — all done inside your web browser.

Please read the data privacy policy first before you start using this app.

The following workflow is recommended:

  1. Click Data to upload your dataset;
  2. Specify the parameters for building models in the Nomogram tab;
  3. Validate and calibrate the built model internally using Internal Validation and Internal Calibration under the Model Validation tab;
  4. Perform Kaplan-Meier analysis and log-rank test for the risk groups of internal calibration using Kaplan-Meier Analysis for Internal Calibration;
  5. (Optional) Validate and calibrate the built model with external datasets using External Validation and External Calibration under the Model Validation tab;
  6. (Optional) Perform Kaplan-Meier analysis and log-rank test for the risk groups of external calibration using Kaplan-Meier Analysis for External Calibration;
  7. (Optional) Compare the models via by Validation and by Calibration under the Model Comparison tab;
  8. (Optional) Predict overall survival probability for new samples based on the built model using the Prediction tab;
  9. Finally, you will be able to download the PDF/HTML/Word report containing the computation results, and the R model object in the Report tab.

Notes:

  • To generate the basic report, please (at least) do step 1, 2, 3, and 4;
  • To generate the external validation report, please do step 1, 2, 5, and 6;
  • To generate the model comparison report, please do step 1 and 7;
  • To download the R model object, please do step 1 and 2.
  • After downloaded the model object, it might be interesting to try the hdnom application maker to make your own nomogram-based online prediction app.

Citation

To cite the hdnom package or the web application in your publications, please click here to view the most recent reference to use.

Feedback

If you have any questions, suggestions, or ideas about the web application, please feel free to let us know:


© 2015 - 2016 Miaozhu Li & Nan Xiao · The hdnom Project · Data Privacy Policy