To understand patterns of geochemical cycling in pristine versus contaminated groundwater ecosystems, pristine shallow groundwater (FW301) and contaminated groundwater (FW106) samples from the Oak Ridge Integrated Field Research Center (OR-IFRC) were sequenced and compared to each other to determine phylogenetic and metabolic difference between the communities. Proteobacteria (e.g., Burkholderia, Pseudomonas) are the most abundant lineages in the pristine community, though a significant proportion (>55%) of the community is composed of poorly characterized low abundance (individually <1%) lineages. The phylogenetic diversity of the pristine community contributed to a broader diversity of metabolic networks than the contaminated community. In addition, the pristine community encodes redundant and mostly complete geochemical cycles distributed over multiple lineages and appears capable of a wide range of metabolic activities. In contrast, many geochemical cycles in the contaminated community appear truncated or minimized due to decreased biodiversity and dominance by Rhodanobacter populations capable of surviving the combination of stresses at the site. These results indicate that the pristine site contains more robust and encodes more functional redundancy than the stressed community, which contributes to more efficient nutrient cycling and adaptability than the stressed community.

In this perspective, we discuss the current challenges and opportunities in multi-omics, a rapidly evolving approach that integrates multiple molecular layers to advance our understanding of complex biological systems. As biomedical research moves toward precision medicine, the ability to correlate genotype, phenotype, and environmental contexts has never been more critical. Multi-omics enhances biomarker discovery and elucidates regulatory networks underlying health and disease. The dominant scientific paradigm for over a century was to take a reductionist approach, studying individual molecular components in isolation or as simplified systems. The advent of omics technologies in the 1990s enabled a systems paradigm, allowing holistic analyses of molecular networks. These early systems studies were constrained by technology and methodology to bulk tissue measurements and single-omics analyses. Recent advances in single-cell and spatial omics, high-throughput proteomics and metabolomics, cloud computing, and artificial intelligence now allow high-resolution, spatially contextualized multi-omics analyses. Despite these gains, challenges in data analysis and interpretation remain, including high dimensionality, missing or incomplete data, multiple batch effects, and method-specific variability. Emerging strategies—such as paired data collection, staged or joint integration, and latent factor or quasi-mediation frameworks—offer promising solutions, positioning multi-omics as a transformative tool for elucidating complex mechanisms and guiding personalized medicine. Continued refinement of these approaches may further enhance the utility of multi-omics for understanding complex biological systems.

The Rhode Island IDeA Network of Biomedical Research Excellence Molecular Informatics Core at the University of Rhode Island Information Technology Services Innovative Learning Technologies develop... Method summary Virtual/augmented reality apps were developed in Unity to teach topics in biomedical science. In-app 3D models were created to provide a scientifically accurate user experience. Apps... Graphical abstract

This manuscript describes the development of a resource module that is part of a learning platform named “NIGMS Sandbox for Cloud-based Learning” https://github.com/NIGMS/NIGMS-Sandbox. The overall genesis of the Sandbox is described in the editorial NIGMS Sandbox at the beginning of this Supplement. This module delivers learning materials on basic principles in biomarker discovery in an interactive format that uses appropriate cloud resources for data access and analyses. In collaboration with Google Cloud, Deloitte Consulting and NIGMS, the Rhode Island INBRE Molecular Informatics Core developed a cloud-based training module for biomarker discovery. The module consists of nine submodules covering various topics on biomarker discovery and assessment and is deployed on the Google Cloud Platform and available for public use through the NIGMS Sandbox. The submodules are written as a series of Jupyter Notebooks utilizing R and Bioconductor for biomarker and omics data analysis. The submodules cover the following topics: 1) introduction to biomarkers; 2) introduction to R data structures; 3) introduction to linear models; 4) introduction to exploratory analysis; 5) rat renal ischemia-reperfusion injury case study; (6) linear and logistic regression for comparison of quantitative biomarkers; 7) exploratory analysis of proteomics IRI data; 8) identification of IRI biomarkers from proteomic data; and 9) machine learning methods for biomarker discovery. Each notebook includes an in-line quiz for self-assessment on the submodule topic and an overview video is available on YouTube (https://www.youtube.com/watch?v=2-Q9Ax8EW84). This manuscript describes the development of a resource module that is part of a learning platform named ``NIGMS Sandbox for Cloud-based Learning'' https://github.com/NIGMS/NIGMS-Sandbox. The overall genesis of the Sandbox is described in the editorial NIGMS Sandbox [1] at the beginning of this Supplement. This module delivers learning materials on the analysis of bulk and single-cell ATAC-seq data in an interactive format that uses appropriate cloud resources for data access and analyses.

UNSTRUCTURED Artificial Intelligence (AI) and immersive simulation offer new opportunities to provide pharmacy students with realistic, standardized patient-care experiences. We developed a three dimensional (3D), AI-enabled virtual patient application designed to simulate a home-based patient encounter and allow students to practice symptom assessment, communication, and clinical reasoning. The application integrates an interactive 3D home environment with an AI-driven virtual patient capable of responding dynamically to student questions and decisions. The application was implemented with pharmacy students during spring 2026 and evaluated for feasibility, student performance, and learner perceptions. Students were also evaluated on their achievement of learning outcomes related to patient assessment. Student performance showed improvement in identifying symptom-management needs, and increased confidence in approaching patients with complex, multiple symptom burdens. Students also reported value in this opportunity to practice patient assessment in a realistic, low-stakes environment. These findings suggest that combining generative AI with an immersive 3D environment can provide a feasible and engaging approach to developing pharmacy students' patient assessment skills.