This review article examines the use of rodent fMRI as a vital tool for understanding brain dysconnectivity in human psychiatric and neurological disorders. While human imaging identifies atypical connectivity patterns, it cannot easily reveal the biological mechanisms or cellular origins of these deficits. By utilizing animal models, researchers can bridge this gap through causal manipulations, such as genetic modifications and neuromodulatory system interventions. The authors highlight the significant spatiotemporal homologies between rodent and human brain networks, which validates the cross-species translation of findings. Furthermore, the text discusses the evolution toward awake imaging and the integration of optogenetics and chemogenetics to decode how microcircuit imbalances manifest as macroscale network dysfunction. Ultimately, these preclinical platforms offer a path to transform general imaging markers into specific mechanistic insights for mental illness.

References:

• Modeling Brain Dysconnectivity in RodentsGozzi, Alessandro et al.Biological Psychiatry, Volume 93, Issue 5, 419 - 429

The paper introduces CellOT, a machine learning framework designed to predict how individual cells respond to biological stimuli, such as drugs or genetic changes. Because measuring a cell's molecular state typically destroys it, researchers cannot observe the same cell before and after a treatment; CellOT overcomes this "pairing problem" by using neural optimal transport to map unperturbed cell populations to perturbed ones. By employing input convex neural networks, the model learns the most likely transition trajectories for heterogeneous cell subpopulations based on a principle of minimal effort. The authors demonstrate that CellOT identifies nuanced, cell-state-specific drug responses and generalizes effectively to previously unseen patients and different species. Compared to existing autoencoder-based methods, this approach more accurately captures the variability and higher-order moments of complex cellular distributions. Ultimately, the research offers a scalable tool for understanding personalized medicine and the mechanistic drivers of cellular evasion in cancer therapies.

References:

• Bunne, C., Stark, S.G., Gut, G. et al. Learning single-cell perturbation responses using neural optimal transport. Nat Methods 20, 1759–1768 (2023). doi.org

The paper details the development of TransBrain, a computational framework designed to bridge the gap between human and animal neuroscientific research. By integrating spatial transcriptomics, structural connectivity, and functional imaging, the tool enables the mapping of molecular and circuit-level data across species. Researchers used this platform to translate mouse optogenetic experiments into predicted human brain activation patterns, linking animal circuit manipulations to human cognitive maps. Furthermore, the study demonstrates TransBrain's clinical utility by aligning structural brain alterations in autism mouse models with individual patient data. This approach successfully identified conserved gene-brain-behavior relationships, particularly concerning the Magel2 gene and intellectual performance. Ultimately, the framework serves as an objective bridge to enhance translational neuroscience and the evaluation of therapeutic targets.

References:

• Huang S, Zhang T, Dong C, et al. TransBrain: A computational framework for translating brain-wide phenotypes between humans and mice[J]. Nature Methods, 2025: 1-12.

The paper introduces cellSTAAR, a novel statistical framework designed to improve the discovery of rare genetic variants within noncoding regions of the human genome. By integrating single-cell epigenetic data with whole-genome sequencing, the method identifies functional associations that are often masked in traditional bulk tissue studies. The tool specifically addresses the difficulty of linking candidate cis-regulatory elements to their target genes by utilizing an ensemble of diverse computational approaches. Applications of the model to large-scale datasets, such as the UK Biobank and TOPMed, demonstrate its superior power in detecting meaningful biological signals for complex traits like lipid levels. Ultimately, cellSTAAR enhances the interpretability of genetic studies by prioritizing the most relevant cell types for specific human diseases.

References:

• Van Buren E, Zhang Y, Li X, et al. cellSTAAR: incorporating single-cell-sequencing-based functional data to boost power in rare variant association testing of noncoding regions[J]. Nature Methods, 2025: 1-12.

Pertpy is a Python-based software framework created to streamline the analysis of large-scale single-cell perturbation experiments, such as genetic modifications or drug treatments. By integrating with the scverse ecosystem, it offers a modular suite of tools for metadata annotation, quality control, and the calculation of specialized perturbation spaces. The platform enhances efficiency through GPU acceleration and provides a unified structure for identifying differential gene expression and shifts in cell type composition. Researchers can use Pertpy to deconvolve complex biological responses, such as distinguishing between viability-dependent and independent drug effects. Furthermore, it facilitates the discovery of multicellular programs and the characterization of previously unannotated gene pathways across diverse disease states. Ultimately, the framework fosters more robust biological insights by connecting experimental readouts with extensive public biological databases and ontologies.

References:

• Heumos L, Ji Y, May L, et al. Pertpy: an end-to-end framework for perturbation analysis[J]. Nature Methods, 2025: 1-10.

The paper introduces SpatialZ, a computational framework designed to overcome current hardware limitations in spatial transcriptomics by reconstructing dense 3D tissue models from sparse 2D sections. Conventional methods often result in data gaps between tissue layers, but this new tool uses interpolation algorithms to generate virtual slices, enabling researchers to map gene expression across entire organs. The software includes modules for in silico sectioning, which allows users to view biological structures from any angle, and 3D continuous rendering for visualizing complex molecular gradients. Validated on mouse brain and human breast cancer data, the technology demonstrates high fidelity in preserving cell-type distributions and anatomical features. Ultimately, this open-source toolkit bridges the gap between planar imaging and comprehensive volume-based biological analysis.

References:

• Lin S, Wang Z, Cui Y, et al. Bridging the dimensional gap from planar spatial transcriptomics to 3D cell atlases[J]. Nature Methods, 2025: 1-13.

This research article introduces IRIS, a novel computational tool designed to identify distinct tissue regions in spatially resolved transcriptomics (SRT) studies. Unlike previous methods that rely on secondary gene expression patterns, IRIS focuses on cell type composition by integrating data from single-cell RNA sequencing (scRNA-seq) references. This approach allows for more accurate and biologically interpretable spatial domain detection across various tissues and species. The authors demonstrate that IRIS is significantly faster and more precise than existing technologies, particularly when analyzing massive datasets like Stereo-seq and 10x Xenium. Ultimately, the method enables researchers to better visualize complex structures such as brain layers, tumor microenvironments, and testicular changes caused by disease.

References:

• Ma, Y., Zhou, X. Accurate and efficient integrative reference-informed spatial domain detection for spatial transcriptomics. Nat Methods 21, 1231–1244 (2024). doi.org

Researchers have developed a suite of new protocols, collectively called FX-Cell, to improve single-cell RNA sequencing in diverse plant species. These methods utilize coagulant fixatives to stabilize cells, allowing them to withstand the high temperatures and mechanical forces needed to break down rigid cell walls. A critical component of the workflow involves a column-based purification step that removes harmful RNases from digestive enzymes, thereby preserving the integrity of the genetic data. This approach enables the study of cryopreserved field samples and reveals accurate wounding responses that traditional methods often distort. By overcoming the technical limitations of standard protoplast isolation, these advancements facilitate the creation of comprehensive Plant Cell Atlases for both model and non-model organisms. Consistent with its versatility, the method successfully captured complex developmental data in tissues as varied as maize anthers and wild rice rhizomes.

References:

• Ming X, Wan M C, Zhang Z D, et al. FX-Cell: a method for single-cell RNA sequencing on difficult-to-digest and cryopreserved plant samples[J]. Nature Methods, 2025: 1-12.

FlashDeconv is a newly developed computational framework designed for high-speed, accurate cell type deconvolution in large-scale spatial transcriptomics. Unlike traditional methods that rely on gene variance—which often ignores rare cells—this tool uses leverage-score importance sampling to preserve distinct biological signals regardless of their abundance. By employing randomized sketching, the system compresses massive datasets to allow for analysis on standard hardware in seconds rather than days. The researchers used this efficiency to identify a "resolution horizon" between 8 and 16 micrometers, demonstrating that coarse data binning can create misleading evidence of cell co-localization. Applied to human ovarian cancer and mouse intestinal tissues, the framework successfully recovered clinical response signatures and uncovered previously hidden Tuft cell niches. Ultimately, FlashDeconv provides a scalable, mathematically grounded solution for mapping the complex cellular architecture of entire tissue atlases.

References:

• Yang C, Chen J, Zhang X. FlashDeconv enables atlas-scale, multi-resolution spatial deconvolution via structure-preserving sketching[J]. bioRxiv, 2025: 2025.12. 22.696108.

This research introduces Deep-STARmap and Deep-RIBOmap, advanced imaging platforms designed for 3D spatial transcriptomics and translatomics within thick tissue samples. Traditional methods often struggle with cellular fragmentation and physical distortion during thin sectioning, but these new techniques utilize hydrogel-based re-embedding and covalent crosslinking to preserve genetic information in larger volumes. The authors demonstrate the versatility of these tools by mapping complex neuronal morphologies in the mouse brain and analyzing the tumor microenvironment of human skin cancer. By maintaining the volumetric context of the tissue, the study reveals intricate cell-to-cell interactions and structural details that are typically lost in 2D analysis. This framework effectively bridges the gap between high-throughput molecular profiling and the physical architecture of intact biological systems.

References:

• Sui X, Lo J A, Luo S, et al. Scalable spatial single-cell transcriptomics and translatomics in 3D thick tissue blocks[J]. Nature Methods, 2025: 1-11.