"Open Problems": New Platform Brings Clarity to the Complexity of Single-Cell Genomics

Single-cell genomics is considered a key technology for understanding health and disease at the cellular level. It enables researchers to analyze individual cells in detail—making it possible to better understand complex disease mechanisms, such as those involved in lung or inflammatory diseases. But with the rapid development of the field, the number of available analysis tools has also exploded: thousands of software tools are now available—but which is best suited for which task?

This is where the new platform “Open Problems” comes in: a collaborative open-source project supported by more than 50 research institutions worldwide, jointly coordinated by Helmholtz Munich and Yale University. The study, recently published in Nature Biotechnology, was led by DZL researchers Dr. Malte Lücken and Prof. Fabian Theis, both based at the DZL site CPC-M.

The goal of Open Problems is to make analysis tools more comparable, enable better quality assessment—and thus foster more efficient and transparent research.

The platform currently provides 81 public datasets, used to benchmark 171 analysis methods across 12 key tasks in single-cell research. Evaluation is based on objective criteria such as accuracy, scalability, and reproducibility. All processes run automatically and according to standardized procedures in the cloud. The platform is open to the international research community: new tasks, methods, and suggestions for improvement can be submitted at any time. "Open Problems lowers the barrier for AI researchers outside the field of biology to get involved in genomics,” says Dr. Malte Lücken, lead author of the study and group leader at the Institute of Computational Biology (ICB) and the Institute of Lung Health and Immunity (LHI) at Helmholtz Munich. “It serves as a blueprint for interdisciplinary innovation.”

More information:
Helmholtz Munich News Article
GitHub Repository

Original publication:
Lücken et al., 2025. Defining and benchmarking open problems in single-cell analysis. Nature Biotechnology. DOI: 10.1038/s41587-025-02694-w