UKbiobank_CVD_project

This repo compiles all the code used for the analysis of 300,000 UK Biobank participants on CVD.

Pre-processing

/This code was used to process data . /preprocessing_data was used to deal with the missingess of the data and removing and adjusting the cases and comorbidities /data_to_impute is used to set the data up for imputation /add_dash add dash score to the data Comorbidities imputed_all_numeric_proper.py is used to actually impute the data

Univariate analysis

UNIVARIATE this code is created for the univariate analysis in which we use glm models and a bonferroni correction to identify varaible most related to outcomes like CVD

Stability Selection

This folder includes 4 scripts, 2 to perform cox regression and the other 2 to perform logistic regression (as a sensitivity analysis) These scripts should be run in order

• The 1_ scripts correspond to the main model of our work, selection and refit were performed: environment and biological variables are set as predictors and CVD as outcome
• The 2_ scripts correspond to the model looking at the indirect effects of the environment on CVD by looking at their effect on the biological variables, selection and refit were performed: environment variables are predictors and biological variables are outcomes.

Prediction Performance

Prediction performance was assed based on the refited models

Creating plots

• this folder includes the codes used in order to plot our figures