MODY-calculator-clinical-paper

This repository contains the code for

• Make predictions for new patients.
• Generating analysis and plots.

Folder structure

1. data: folder containing the data used for the analysis (empty as data cannot be shared)

2. figures: folder containing figures used for the manuscript

3. model_development: R code and model fitting objects (several models are included based on the different models fitted for the article)

4. model_predictions: Vectors containing MODY probability estimations for all of the datasets used (_full.rds represents the full posterior dataset, its omission represents the summary results (2.5%, mean, 97.5%))

5. new_data_predictions: R code + data + Excel document for making predictions in new datasets

6. bootstrap_riley_simulation: R code for a simulation study on stability of model predictions based on bootstrapped datasets.

File explanation

• discrimination_curves.R: this file contains the R code for calculating and plotting ROC and Precision-recall curves.
• bootstrap_validation_analysis.R: this file contains the R code for making calibration plots of bootstrapped datasets.
• model_convergence.R: this file contains the R code for checking model convergence in the main models used for this analysis.
• calibration_plots.R: this file contains the R code for creating the calibration plots for the datasets used.
• CPRD_analysis.R: this file contains the R code for plotting density plots for probability of MODY.
• antibody_sensitivity_analysis.R: this contains the R code for plotting probabilities of MODY based on different antibody information availability.