Working Groups

Bring your questions and your data—NSCORE will connect you with a multidisciplinary team to integrate and analyze them. Working groups form the core of NSCORE's synthesis science. Each is typically composed of up to eight scientists, mixing biologists and computer scientists at a range of career stages, and is supported for up to two years. During the application stage, group leaders work with NSCORE to articulate the questions and goals of their synthesis science. Successful proposals receive a range of support, including data scientist and student time, cyberinfrastructure and GPU access, a travel budget, and facilitation and administrative assistance. Leaders also undergo leadership and team science training to help them guide cross-disciplinary collaboration. Groups meet at NSCORE once or twice a year and virtually each month, working toward peer-reviewed publications of the highest quality, new tools and datasets for synthesis science, and grant proposals that seed new collaborations and empirical projects. Visit Participate to learn how to apply.

Current Working Groups

Behavioral flexibility. Animal movement, and the fine-scale behavior embedded within it, is one of the most sensitive and immediate expressions of organismal response to a changing environment, yet it remains one of the hardest dimensions of that response to measure consistently across species, ecosystems, and continents. Accelerometers are now deployed as a near-default complement to GPS tracking on thousands of individual organisms worldwide, and they already record fine-scale kinematic information continuously. Yet, the resulting archives have accumulated under a patchwork of sampling regimes, tag designs, and single-study aims. We also have growing archives of remote video of animal behavior, which can be analyzed in similar ways. Using billions of accelerometer records, combined with video analysis and advances in machine learning models, this group will ask whether the fine-scale behavioral signals encoded in acceleration, together with concurrent GPS records, can improve models of animal movement and provide biodiversity metrics useful for conservation and policy. This working group will be co-led by Roland Kays (North Carolina State University) and Alex Jordan (Max Planck Institute of Animal Behavior).

Molecular plasticity. Alternative splicing expands transcriptome diversity and gives organisms a flexible layer of gene regulation, letting genotype translate into phenotype in response to a changing environment. By refining the training of genome language models with a large public compendium of 20,000+ Arabidopsis thaliana transcriptomes that cover a range of experimentally applied perturbations, alongside several existing collections of transcriptomic variation from wild A. thaliana populations, this group aims to distinguish mechanisms of molecular plasticity that are shared across many kinds of stress from those unique to a specific pressure. In addition, the group will test how well lab-induced responses actually reflect what happens in wild populations. The result will generate a more integrative, mechanistic picture of how plants build organismal resilience, and demonstrate the application of genome language models to a broader range of organisms. This working group will be co-led by Joy Bergelson (New York University) and Itsik Pe’er (Columbia University).

Morphological Plasticity. Why do some forests withstand climatic extremes while others undergo lasting structural change? Part of the answer may lie in the trees themselves. Species differ in maximum attainable height, crown architecture, wood density, shade tolerance, and hydraulic strategy, and in how far they can adjust their growth and form under stress. This group will use deep learning to test whether the traits of individual trees, species identity, and community composition predict how the three-dimensional structure of a forest resists and recovers after droughts, heat waves, and storms worldwide. Global vegetation surveys and long-term forest inventories will be paired with repeat airborne laser scanning and spaceborne lidar, as well as with trait data from existing databases and herbarium collections. AI-based image analysis will extract morphological traits directly from digitized specimens. AI models trained on these data will link traits and composition to canopy architecture to track structural change responses in forests to documented climate extreme events. The group will produce a global data synthesis and a predictive framework of forest resilience that scales from single trees to whole forests. This working group will be co-led by Sheng-Feng Shen (Academia Sinica), Mukund Rao (Columbia University), Emily Sessa (New York Botanical Garden), and Brad Oberle (New York Botanical Garden).