Applying Big Data Analytics to Answer More Questions About ALS

Share this post:

Each year, more than 5,600 Americans are diagnosed with Amyotrophic Lateral Sclerosis (ALS), also called Lou Gehrig’s Disease. This diagnosis usually prompts more questions from patients and family members about longevity and quality of life than what physicians can immediately answer.

Unfortunately, the same is true for ALS research, which historically has been hindered by limited access to patient-level data and sophisticated computational tools. At Prize4Life, a nonprofit founded in 2006, we’re dedicated to accelerating the discovery of treatments and a cure for ALS – and we believe that data, and easy access to it, holds the key.

That’s why we recently teamed with IBM, DREAM and Sage Bionetworks, leaders in big data and analytics, crowdsourced challenges and patient research, to create and conduct a clinical challenge to help answer some of these patients’ and families’ questions.

The result: the DREAM ALS Stratification Prize4Life Challenge. This crowdsourced competition combined the power of clinical trial data with the developer community and an advanced Linux computing system to gain meaningful insights into ALS.

Here’s how the Challenge worked:

• More than 250 individual or team participants had access to our Pooled Resource Open-Access Clinical Trial database of over 12 million data points from 10,000 de-identified patient data records – the world’s largest such collection. This allows them to analyze demographics, clinical information, family history, lab data and more.
• Using this data, the teams developed predictive models to identify clinical features of ALS pathology and sub-groups of ALS patients that are affiliated with disease progression and survival.
• To run these data-intensive models, IBM made its IBM LinuxONE system available to each of the teams via the cloud. LinuxONE, the world’s fastest and most secure Linux system, provides the security and speed needed to run analytics up to seven times faster than commodity servers, and scale to support such a large competition.
• Sage Bionetworks and DREAM created a cloud-based Challenge environment where participants could access and analyze the data, submit their quantitative solutions to a leaderboard for scoring (via the IBM Cloud) against a hidden validation data set and share their ideas, code, and results with others in the Challenge.

Through analyzing the Challenge data, the hope was that Challenge participants could identify patterns of clinical features that can help differentiate why certain groups of patients live longer than others. With accurate stratification models like these, doctors can more effectively develop potential treatments for ALS.

The winning teams were determined by running models on LinuxONE to gauge their accuracy at predicting survivability. The winners were announced on November 5, and the outcomes already suggest leads for new clinical features related to disease progression and survival. These models will now be made available to the scientific community to enable further ALS research.

When we look at the massive challenge that lies ahead in the quest to find a cure for ALS, we see opportunity in the combined resources of science, technology and medicine. With partners like IBM, DREAM and Sage Bionetworks, we are convinced that new ways of analyzing and modeling patient data can help uncover patterns that can truly mean the world to those who are affected by this disease.
_______________________________

Reach Neta Zach at nzach@prize4life. org