DREAM Challenges
October 2, 2014


Open Challenges

  • RA2 DREAM Challenge

    RA2 DREAM Challenge: Automated Scoring of Radiographic Joint Damage

    Launch: November 4, 2019 (Now open)
    Final Submissions: May 31, 2020
    Winners Announced: July 31, 2020

    Rheumatoid Arthritis (RA) is a debilitating disease that causes joint damage in the hands and feet due to inflammation. Accurate measuring of joint damage and progression of disease is essential in assessing the severity of disease, along with being able to monitor patient response to treatment. Scoring of X-ray images is tedious and there can be a great deal of variability between rheumatologists when applying scoring methods. We will leverage nearly 1000 images across a range of disease severities to challenge participants to develop an automated scoring algorithm with the goal of making clinical scoring fast and consistent; this will aid rheumatologists in making the best decisions for patients.


  • Allen Institute Cell Lineage Reconstruction DREAM Challenge

    Allen Institute Cell Lineage Reconstruction DREAM Challenge

    Registration Opens: September, 2019
    Launch: Mid-October, 2019 (Now Open)
    Final Submissions: January, 2020
    Winners Announced: March, 2020

    CRISPR-based molecular tools allow the reconstruction of cell lineages based on the phylogenetic analysis of DNA mutations induced by CRISPR during development. The challenge - a partnership between The Allen Institute and DREAM - will comprise 3 subchallenges focused on reconstructing cell lineage trees of different sizes and nature. Through the usage for training purposes of experimental and in silico generated data sets, the goal of this challenge is to mobilize a larger community for evaluating new optimal tree-building methods.

  • The EHR DREAM Challenge

    The EHR DREAM Challenge

    Registration OPEN
    Launch October 9, 2019
    Final submissions January 9, 2020
    Results announced February 9, 2020

    The goals of the EHR DREAM Challenge are to prospectively predict patient health outcomes in order to 1) lower the barriers to piloting innovative machine learning and data science methods in healthcare, 2) establish clinically relevant prediction benchmarks and evaluation metrics, and 3) minimize the distance between model developers and the clinic to reduce the time for model implementation.


  • Single Cell Signaling in Breast Cancer

    Single Cell Signaling in Breast Cancer

    Launch: August 2, 2019 (now open)
    Webinar: August 20, 2019

    This challenges uses genomic, transcriptomic, and proteomic measurements to evaluate signaling responses of 67 well-characterized breast cell lines to drug treatment at the single-cell level to better understand single-cell signaling and signaling modeling.

  • Tumor Deconvolution DREAM Challenge

    Tumor Deconvolution DREAM Challenge

    Launch:  June 3, 2019 (Challenge Now Open!)
    Webinar: June 5, 2019

    The goal of this Challenge is to evaluate the ability of computational methods to deconvolve bulk expression data, reflecting a mixture of cell types, into individual immune components.

  • Preterm Birth Prediction: Transciptomics DREAM Challenge

    Preterm Birth Prediction: Transciptomics DREAM Challenge

    Sub-challenge 1: May 4-Aug 15, 2019
    Top performers announced:  September 13, 2019
    Sub-challenge 2: Aug 15, 2019 - Dec 5, 2019 (Register now)
    Top performers announced:  January 3, 2020

    Using whole blood gene expression data collected from pregnant women, participants will develop models to predict gestational age at blood draw and risk of preterm birth.