DreamTools aims at sharing code used in the scoring of DREAM challenges.

The main goals are to provide:

1. scoring functions for the Dream challenges for end-users via the dreamtools-scoring standalone application.
2. a place for developers involved in the dream challenges to share code

Code related to aggregation, leaderboards, or more complex analysis are not guaranteed to be found in dreamtools even though it may be in some challenges (e.g. D8C1). Some functionalities may be restricted with some access to synapse. Some challenges will require to download public data sets from Synapse (in which case you will need to register).

DreamTools depends on a few libraries such as Pandas, Numpy, Matplotlib. They should be automatically installed with dreamtools using pip executable:

pip install dreamtools

From source, download the source code and install the package with Python:

git clone git@github.com:dreamtools/dreamtools.git
cd dreamtools
sudo python setup.py install

For users, dreamtools provide one executable called dreamtools-scoring, which should be installed automatically when installing dreamtools. Knowing the name of the challenge (and possibly sub-challenge), it works as follows:

dreamtools-scoring --challenge d8c1 --sub-challenge sc1a --submission example.zip

It prints some information and the score of the submision for instance for the example above:

Solution for alphabeta-Network.zip in challenge d8c1 (sub-challenge sc1a) is :
AUROC: 0.803628919403
Rank LB: 1

Dreamtools software does not include all scoring functions but more will be implemented in the future. Here is the list of challenges available

1. DREAM9.5

   • D9dot5C1 olfaction challenge

2. DREAM8

   • D8C1 (HPN breat cancer) sub challenges named sc1a, sc1b, sc2a, sc2b. See Synapse page D8C1 for details
   • D8C2 (Toxoxicology) sub challenges named sc1, sc2. See Synapse page D8C2 for details
   • D8C3 coming soon

3. DREAM7

   • D7C1 (Network Parameter Estimation)

4. DREAM6

   • D6C1 (Network Parameter Estimation) WONT be included (see D7C1 instead)

5. DREAM5

   • D5C2 (transcription factor)

6. DREAM4

   • D4C1 (Peptide Recognition Domain (PRD) Specificity Prediction)
   • DREAM4 In Silico Network Challenge
   • D4C3 (Predictive Signaling Network Modeling)

7. DREAM3

   • D3C1 (Signaling Cascade Challenge)
   • D3C2 (Signaling Response Prediction)

Format of the template for each challenge should be found in the README of each subchallenge. For instance, for Dream8 Challenge1, please see ./dreamtools/dream8/D8C1/README.rst

Please fill bug report in https://github.com/dreamtools/dreamtools/issues

Please join https://github.com/dreamtools/dreamtools

Please see doc/source/developers.rst

Please see doc/source/credits.rst