Skip to content
Gene-disease druggability prediction collaboration [GSK + B2SLab]
PostScript TeX R Mathematica
Branch: master
Clone or download
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
00_metadata Added packrat summary table Aug 1, 2018
00_rawdata Added network-based features for reproducibility May 14, 2019
01_data Analysis 5.1 (preliminary) Oct 2, 2017
02_performance Analysis 5.1 (preliminary) Oct 2, 2017
03_data Analysis 5.3 Oct 20, 2017
03_performance Analysis 5.3 Oct 20, 2017
04_topology Run 5.3 bis Oct 21, 2017
05_mashup Added packrat and other files for reproducibility Dec 27, 2017
10_data All diseases + logistic models Nov 27, 2017
11_topology Disease clustering Nov 28, 2017
12_performance Predictions using genetic scores as well Dec 11, 2017
13_complexes quick fix Dec 21, 2017
20_data Added genetic scores histogram Aug 2, 2018
21_topology Exported data frames with topological properties Mar 29, 2018
22_performance CV runs on STRING Feb 6, 2018
23_boxplots Models on all networks Feb 9, 2018
23_contrasts Models on all networks Feb 9, 2018
23_models Models on all networks Feb 9, 2018
40_data Added OT network stats & complex stats. Updated complex stats, empty … Mar 27, 2018
42_performance CV runs OmniPath Feb 6, 2018
43_boxplots Models on all networks Feb 9, 2018
43_contrasts Models on all networks Feb 9, 2018
43_models Models on all networks Feb 9, 2018
45_mashup Generated dataset, network and features for OmniPath. Added biomaRt l… Jan 30, 2018
63_boxplots Modified method ranking figure Aug 23, 2018
63_models Added plot on predictions and updated text files (probably emmeans vs… Aug 1, 2018
packrat Finally installed the GGally dependence Jul 30, 2018
.Renviron
.Rprofile Added packrat and other files for reproducibility Dec 27, 2017
.gitignore Added packrat sources - experimental as they are large Dec 27, 2017
00_packrat_table.R Added packrat summary table Aug 1, 2018
01_preprocessing.Rmd Analysis 5.1 (preliminary) Oct 2, 2017
02_diffusion_scores.Rmd Analysis 5.1 (preliminary) Oct 2, 2017
03_config.R First descriptive statistics on complex data Dec 4, 2017
03_multiple_disease.Rmd
03_preprocessing.Rmd Added scripts for analysing 4 diseases Oct 11, 2017
04_positives_analysis.Rmd Run 5.3 bis Oct 21, 2017
05_mashup.m All diseases + logistic models Nov 27, 2017
05_mashup_features.Rmd All diseases + logistic models Nov 27, 2017
10_preprocessing.Rmd Now genes are not filtered if no drugs or genetic association is known Nov 15, 2017
11_positives_analysis.Rmd Disease clustering Nov 28, 2017
11_upgma.R Disease clustering Nov 28, 2017
12_multiple_disease.Rmd Predictions using genetic scores as well Dec 11, 2017
13_complexes.Rmd Added simulated CV folds Dec 5, 2017
13_pilot_cv_schemes.Rmd quick fix Dec 21, 2017
20_config.R Added fold imbalance plot Apr 3, 2018
20_preprocessing.Rmd Added genetic scores histogram Aug 2, 2018
21_positives_analysis.Rmd Exported data frames with topological properties Mar 29, 2018
22_performance.Rmd CV runs on STRING Feb 6, 2018
23_models.Rmd Models on all networks Feb 9, 2018
40_config.R Added OT network stats & complex stats. Updated complex stats, empty … Mar 27, 2018
40_preprocessing.Rmd Added OT network stats & complex stats. Updated complex stats, empty … Mar 27, 2018
42_performance.Rmd CV runs OmniPath Feb 6, 2018
43_models.Rmd Models on all networks Feb 9, 2018
45_mashup.m Generated dataset, network and features for OmniPath. Added biomaRt l… Jan 30, 2018
60_abbreviations.R Added topology analysis on STRING. Abbreviations are now in a config … Mar 28, 2018
60_config.R Added new boxplots by disease/method Mar 29, 2018
60_palette25.txt Added new boxplots by disease/method Mar 29, 2018
63_models.Rmd Modified method ranking figure Aug 23, 2018
LICENSE Create LICENSE Oct 23, 2018
README.md
config.R exploratory analysis Sep 18, 2017
genease.Rproj exploratory analysis Sep 18, 2017

README.md

Introduction

The genedise project aims at finding druggable genes for a specific disease based on previously essayed targets. Whether these targets were successful or not is not the primary concern - the fact that there was enough evidence to try them is enough for us. In this way, we aim at mimicking the time-consuming task of proposing new reasonable targets.

The suggestion of new disease genes uses data from OpenTargets as seed gene lists and the STRING protein-protein interaction network to infer new genes.

The project is almost entirely coded using R. Some Matlab code has been necessary to include state of the art approaches.

Structure

The files and directories of this project are proceded by a number that indicates the chronological order of their execution. Scripts are stored in Rmd files. Their outputs are saved in folders sharing their prefix. The most relevant prefixes are:

  • 2X_: analysis on the STRING network
  • 4X_: analysis on the OmniPath network
  • 6X_: plots and models combining both networks (depends on the execution of the 2X an 4X scripts)

Reproducibility

Metadata files

The output of sessionInfo() is always stored in the directory 00_metadata to keep track of the package versions.

Configuration files

There are configuration files, such as 03_config.R, that contain a comprehensive amount of parameters, paths and file names. Generally, these parameters are sourced instead of being hardcoded in the scripts.

Package management

The project has package version control through packrat to ease portability between machines.

External files

Almost all the files in the project are included in the git repository at the moment. Exceptions:

  • STRING database files
  • Network kernel(s)

The route of these files (Sergi's machines) can be found in the config files.

Other

There are several set.seed calls throughout the code. Intermediate results are saved when the space required is not prohibitive.

Workflow

Data preprocessing

  • Check OpenTargets data sanity
  • Choose network: compromise between coverage and size
  • Compute and store graph kernel on chosen network
  • Save cleaned data, mapped to the network of choice

Topology analysis

  • Characterisation of disease genes in terms of network properties
  • Within-disease study
  • Between-disease study

Performance

  1. Load configuration files

  2. Load dataset

  3. Load network data

  4. Build CV folds

  5. Define functions for prediction

  6. Define performance metrics

  7. For each disease,input_type,fold

    1. Define train and validation
    2. Predict for every method using train
    3. Compute performance metrics
    4. Write to disk
  8. Plot metrics

  9. Build statistical models for comparing methods

System requirements

Hardware

The runs have been executed on the following hardware from the UPC:

Code profiling

Running the script is barely possible with 16GB of RAM. We recommend using 32GB to avoid spikes with swapping.

For reference, executing all the diseases under a single repeated CV scheme (25 repetitions, 3 folds per repetition) on eko takes one week. Likewise, sun is twice as fast. The code is a mixture between serial and parallel executions because not all the methods run in parallel.

On the other hand, the computationally intensive code was run on a torque-based cluster, but the parallel R package -part of the R base- was unable to clean up the child processes. This led to memory exhaustion and proved to be infeasible. Alternatives to tackle this while keeping reproducibility might be added in the future.

You can’t perform that action at this time.