Skip to content

Random Forest classifier designed to predict pairs of human genes capable to causing a digenic disease when carrying rare variants simultaneously. DiGePred has been trained using digenic pairs from DIDA and non-digenic pairs from unaffected relatives of individuals with rare undiagnosed disease.

Notifications You must be signed in to change notification settings

oligogenic/DiGePred

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DiGePred

Random Forest classifier designed to predict pairs of human genes capable to causing a digenic disease when carrying rare variants simultaneously.

DiGePred has been trained using digenic pairs from DIDA (Gazzo et al., 2016; http://dida.ibsquare.be) and non-digenic pairs from various sources.

unaffected: non-digenic pairs derived from unaffected relatives of individuals with rare undiagnosed diseases, as part of the Undiagnosed Diseases Network. permuted: non-digenic pairs derived from permutations of genes from digenic pairs in DIDA. random: non-digenic pairs generated randomly. matched: non-digenic pairs chosen to match the distribution of features of digenic pairs. unaffected no gene overlap: unaffected non-digenic pairs chosen such that no genes were common between the training and held-out test sets.
random no gene overlap: random non-digenic pairs chosen such that no genes were common between the training and held-out test sets.

The positive (digenic) and negative (non-digenic) gene pairs used to train and test DiGePred are provide in the folders "positives" and "negatives".

All trained DiGePred models are provided in the folder "models". "Unaffected-no-gene-overlap" was the final and best performing model.

Scripts to train models and test DiGePred performance have been provided in "scripts".

DiGePred_train_all_sets.py: script to train the classifier and test performance during training. DiGePred_held_out_test_performance.py: script to get predictions on held-out test and measure performance using ROC and PR curves. Get_DiGePred_scores_user_input_genes_or_pairs.py: script where user may provide an input file (.txt) containing either list of genes (-g) or list of gene pairs (-p), along with model of choice (-m) and job name (-n). It generates DiGePred results CSV with feature values and DiGePred predictions, based on model of choice.

DiGePred has been run on all human gene pairs, based on all genes from HGNC. The scores are available here: https://vanderbilt.box.com/s/n1nzdyj8i5fa55vultyq4xn6rsp792a7 https://vanderbilt.box.com/s/459ethsqv339nqiarhm0j227jdjb0whq https://vanderbilt.box.com/s/acdqvjuihj3932c6msi5py82rvr5kam3 https://vanderbilt.box.com/s/kb3vzubfxjcjtxt8x0y1vytu59x8r8no

To fetch a pre-computed DiGePred score, grep 'geneA,geneB' on the files mentioned above. (genes should be in alphabetical order)

A website is also available for the user to access DiGePred scores for all human gene pairs (http://www.meilerlab.org/index.php/servers/show?s_id=28).

Requirements networkx v1.9 sklearn

About

Random Forest classifier designed to predict pairs of human genes capable to causing a digenic disease when carrying rare variants simultaneously. DiGePred has been trained using digenic pairs from DIDA and non-digenic pairs from unaffected relatives of individuals with rare undiagnosed disease.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%