Skip to content

PPMI Data Challenge 2016: A Decision Support System for Parkinson’s Clinical Trials Design

Notifications You must be signed in to change notification settings

brian-j-smith/PPMIChallenge2016

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 

Repository files navigation

PPMI DATA CHALLENGE 2016

The Parkinson's Progression Markers Initiative (PPMI) has generated a comprehensive, standardized, longitudinal set of clinical, biological and imaging data unique to the Parkinson's disease (PD) field and ripe for novel and innovative exploration. Study sponsor The Michael J. Fox Foundation (MJFF) seeks to stimulate analysis of this robust dataset while offering an opportunity for cross-disciplinary collaboration. By casting an open call to a rapidly expanding network of computational scientists, data scientists and neuroscientists, MJFF aims to gather new insights into PD diagnosis and progression. In an effort to assist analysts who may be unfamiliar with Parkinson's disease, MJFF has established a team of subject matter experts to collaborate with applicants as they analyze the PPMI dataset. These PD experts will assist in understanding the rationale for the challenge questions and the nuances of the data set components.

CHALLENGE TO THE DATA SCIENCE COMMUNITY

Develop a quantitative model to answer one of the following questions:

  1. What factors at baseline predict clinical progression?
  2. What are the sub–types of Parkinson's disease?

FORMAT

The PPMI Data Challenge offers $50,000 in total prizes, awarding a $25,000 cash prize to the most comprehensive and applicable answer to each question. GE Healthcare, a PPMI industry partner, and MJFF each are providing $25,000 for the prizes.

Parkinson's is highly variable, with age of onset, rate of progression, and type and severity of symptoms different across the 5 million worldwide living with the disease. Identifying models for prognosis and sub-typing would aid in subject selection for clinical studies and design of trials toward novel therapies.

Applicants will submit a Submission Template (five pages maximum) containing:

  • Analytical methods used for addressing the specific challenge question
  • Statistical analysis or model detailing the predictive accuracy
  • Impact of the analysis
  • Additional supporting figures, images and code (optional)

The submission template is intentionally broad, as MJFF hopes to encourage creativity in analyzing the data and answering the challenge questions.

Consistent with the spirit of PPMI, any algorithms, inventions or discoveries made using the PPMI dataset are subject to the PPMI Data Use Agreement. We expect these algorithms to be shared with MJFF and the research community. Organizations with existing background IP may maintain their IP, while using the PPMI dataset as a tool for validation.

REVIEW PROCESS

The Data Challenge Review Committee, comprising MJFF staff and a team of external reviewers, will convene to review applications and determine the winner of each challenge question. The Review Committee will include experts in statistics, data science and Parkinson's disease research and will judge submissions based on the following criteria:

  • Relevance/suitability of answering questions
  • Methodology of analysis
  • Predictive accuracy of statstical model and algorithms
  • Impact, relevancy and usability in PD research

CHALLENGE TIMELINE

Date Event
June 7, 2016 Challenge Launch
June 21, 2016 Informational Webinar
September 7, 2016 Submissions Due
November 2016 Announce Prize

LINKS

About

PPMI Data Challenge 2016: A Decision Support System for Parkinson’s Clinical Trials Design

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages