The Future of Neurodegenerative Research Depends on Data Interoperability

Neurodegenerative diseases like ALS, Alzheimer’s, and Parkinson’s are not single diseases—they are constellations of biological subtypes hiding behind shared clinical symptoms. A diagnosis of ALS can result from any number of distinct molecular pathways that converge to lead to motor neuron degeneration. To truly understand root causes across patient subpopulations, researchers need access to data at a scale that traditional biomedical research has rarely achieved. Thousands—if not tens of thousands—of patient datasets spanning genomics, transcriptomics, imaging, clinical phenotypes, and longitudinal outcomes are required to disentangle signal from noise.

For decades, the field has operated in silos—individual labs, institutions, and even countries working independently. This model made sense in an earlier era, but today it represents a structural bottleneck. The result is a global patchwork of datasets big and small. Many of these datasets, individually, are scientifically insufficient, but collectively, they could be transformative. Enabling interoperability of these accumulating datasets is a key infrastructure challenge of our time.

Recognizing this, the field has begun to shift. Large-scale initiatives have emerged to aggregate datasets into centralized, cloud-based repositories. These efforts have required overcoming significant cultural and institutional barriers around data ownership, privacy, and competition. They represent real progress and are already enabling more robust analyses than ever before. A really nice example of this is the recent launch of the AMP-ALS Knowledge portal, hosting a collection of large, high quality clinical datasets from across academia, non-profits, and biotech and including Target ALS Global Natural History Study Data.

But aggregation is not the same as interoperability.

Physically moving data into a single repository is a costly, time-consuming, and often fragile solution. Data sharing agreements can take anywhere from 3 months to a year to execute. Transferring large datasets requires substantial infrastructure and introduces opportunities for error, data loss, or corruption of metadata. Privacy concerns—especially under frameworks like HIPAA in the United States and GDPR in Europe—add further complexity, particularly for cross-border data sharing. Even when data is successfully centralized, inconsistencies in formatting, terminology, and measurement units can limit its usability. Recently, Target ALS joined Michael J Fox Foundation and a number of other private funders of big datasets to discuss FAIR data principles – Findability, Accessibility, Interoperability, and Reusability. The outcome of the workshop was recently published in the Journal of Parkinson’s Disease and led to continued collaboration to develop mapping approaches for harmonizing our data with other major datasets to support federated learning across the datasets.

To enhance data findability, accessibility and interoperability, the Target ALS Data Engine is powered by DNAstack’s Omics AI platform, which allows researchers to discover and access large genomic and multi-omic datasets across rare and neurodegenerative diseases through a single, unified standards compliant interface, without having to move sensitive data. Federated systems are often easier to implement prospectively—when datasets are being generated with interoperability in mind—than retrospectively, where legacy datasets exist in incompatible formats across disparate systems. In this scenario, scientists are turning to federated learning approaches, where machine learning algorithms are sent to distributed datasets, trained securely, and then aggregated into global models without moving raw data. Groups like Flower.ai develop industry-grade applications to provide the infrastructure for federating learning.

These approaches represent a meaningful step forward. They reduce friction, enhance privacy, and make it possible to leverage data at scale in ways that were previously impractical. While this has been a brief deep dive into mechanics of data sharing and interoperability, it is important to highlight that the meaningful output of these analyses, traditional or ML-based, rests on the quality of the data used. Aligned with this goal, Target ALS’s ALS Global Research Initiative is designed to generate comprehensive, multi-modal datasets from people from all backgrounds, ethnicities and geographic regions. These data can be found in the Target ALS Data Engine and will be continually updated with a goal of reaching 6000 participants over the next 5 years. We look forward to the new breakthroughs that will come from these data and interoperability with other large datasets across neurodegenerative disease.

Related News

Q1 2026 Publications from our Innovation Ecosystem

March 31, 2026

Q1 2026 Research we Enable Update

March 31, 2026

Q1 2026 Grants & Funding Opportunities Update

March 31, 2026

Q1 2026 Research Spotlights

March 31, 2026