Summary

Single-cell sequencing technologies have revolutionized our understanding of the heterogeneity and dynamics of cells and tissues. However, single-cell data analysis faces challenges such as high dimensionality, sparsity, noise, and limited ground truth.

In this 3rd installment of the Open Problems in Single-Cell Analysis competitions at NeurIPS, we challenge competitors to develop algorithms capable of predicting single-cell perturbation responses across experimental conditions and cell types.

We will provide a new benchmark dataset of human peripheral blood cells under chemical perturbations, simulating drug discovery experiments. The objective is to develop methods that can generalize to unseen perturbations and cell types, enabling scientists to overcome the practical and economic limitations of single-cell perturbation studies.

The goal of this competition is to leverage advances in representation learning—especially self-supervised, multi-view, and transfer learning—to unlock new capabilities bridging data science, machine learning, and computational biology. We hope this effort will continue to foster collaboration between the computational biology and machine learning communities to advance the development of algorithms for biomedical data.

Sponsors

Cellarity is redefining drug discovery by targeting cell behaviors as opposed to individual proteins.

CZI leverages technology and collaboration to accelerate progress in science, education and community work.

Organizers

In alphabetical order:

• Daniel Burkhardt (Primary contact) is a Machine Learning Scientist at Cellarity, a biotechnology company in Boston. He is a core organizer of the Open Problems in Single-Cell Analysis project. He completed his PhD in Genetics at Yale University, specializing in machine learning under Smita Krishnaswamy. His dissertation focused on modeling experimental perturbations and biological heterogeneity in single-cell datasets. Daniel is also the core organizer of the Machine Learning for Single Cell Analysis workshop, offered bi-annually since 2019 by the Krishnaswamy Lab at Yale.

• Robrecht Cannoodt is a computer science engineer at Data Intuitive. He completed his PhD in Saeys Lab at Ghent University, focusing mainly on unsupervised learning in single-cell omics, particularly trajectory inference. In OpenProblems, Robrecht oversees infrastructure development for building collaborative and modular pipelines using Nextflow and Viash.

• Scott Gigante is an Associate Staff Machine Learning Scientist at Immunai, a biotechnology company in New York. He is a core organizer of the Open Problems in Single-Cell Analysis project. He completed his PhD in Computational Biology at Yale University, specializing in machine learning under Smita Krishnaswamy and Ronald Coifman. Scott is also the core maintainer of the Open Problems for Single Cell Analysis living benchmark, focusing on maximizing transferability of solutions from competitions to benefit the broader community.

• Christopher Lance is a PhD candidate in the Machine Learning Group of Prof. Fabian Theis at the Helmholtz Center Munich. He graduated in Molecular Biotechnology at the University of Heidelberg, focusing on bioinformatics. In prior research projects at the Sanger Institute, UK, and EMBL Heidelberg, he worked on association studies of rare genetic variants with plasma proteomics and leveraging metabolomics data to characterize drug perturbations on bacterial metabolism. Christopher joined the OpenProblems team during the first NeurIPS competition in 2021 on multimodal single-cell data integration as one of the main data analysts and was in charge of follow-up assessments of competition results, published in PMLR. He is now focusing on best practices for analyzing multimodal single-cell data and integrating multiple modalities. Christopher will contribute to analyzing this year's data and evaluating baseline methods and metrics.

• Malte Lücken is a Principal Investigator at Helmholtz Munich and a core organizer of the Open Problems in Single-Cell Analysis project. His research focuses on building single-cell reference atlases, benchmarking computational methods for single-cell analysis, and investigating how environmental stimuli and natural variation manifest at the single-cell level. Malte has organized hackathons for the Single Cell Omics Germany network, Helmholtz Munich, and the Open Problems project. He will supervise the competition and shape datasets and metrics to answer key questions in perturbation biology.

• Angela Pisco is the Director for Computational Biology at insitro, a biotechnology company in South San Francisco, and a core organizer of the Open Problems in Single-Cell Analysis project. Her main research interests include single-cell genomics, focusing on building single-cell atlases to understand health and disease. Angela’s team is passionate about extracting meaningful information from biomedical datasets to improve disease understanding and drug development.