Raw data is available in a private Kaggle repository. Access may be requested.
All results are available here.
Reproducibility usage:
-
All representations (traits) generated through Deep Learning algorithms can be extracted by running 'run_deep_models.py'. It will download the raw data and do all the cleaning processes and feature engineering. Then it will fit all the models and extract the representations.
-
With the representations: GWAS, heritability and other statistics used can be replicated by runing 'GWAS_R2_H2.R' in src/model/.
-
Reverse GWAS can be run with 'Reverse_GWAS.py' in src/model/.
-
With representations and statistics: All generated images can be replicated using 'Plot_maker.ipynb' and 'Plot_maker.Rmd'.