A foundation model is trained on a broad body of data to learn general patterns and can then be adapted to many tasks.
A new technique called abcFISH, a multiplex single-molecule fluorescence in situ hybridization strategy, enables spatial imaging of circular RNAs.
Spatial multi-omic technologies measure multiple molecular layers within intact tissues. Two central challenges are vertical integration, which combines different molecular modalities, and horizontal integration, which aligns cell types across studies. Here we show domain invariant representation through adaptive calibration (DIRAC), a geometric deep learning model that unifies these tasks in a generalized domain adaptation framework. DIRAC learns task-discriminative features that are insensitive to domain shifts caused by batch effects, omic modalities or study sites, enabling a shared embedding space and automated cell-type annotation through label transfer from reference datasets. Across sequencing-based and imaging-based datasets, DIRAC identified a new microvascular proliferation domain in glioblastoma, corrected cortical-layer annotations in severe Alzheimer’s disease, resolved subtle structures in mouse embryos and remained robust to missing data and unseen cell types. We generated mouse spleen and thymus Stereo-CITE-seq data and used DIRAC to reveal thymic T cell migration and splenic immune cell interactions. DIRAC is fast and scales to millions of cells. DIRAC is a geometric deep learning model that enables both vertical and horizontal integration and annotation of spatial multi-omic data.
We argue that cell number must be treated as a core experimental parameter to elevate microscopy beyond simple trend detection towards robust quantitative measurements capable of revealing the full complexity of biological systems. We demonstrate that, while central tendencies are readily captured, resolving a distribution’s underlying shape and its rare events necessitates a hundredfold increase in cell number, often rendering high magnification microscopy statistically underpowered.
Capturing the fine details of behavior when animals perform complex tasks, such as prey pursuit and capture, in natural environments is a challenging endeavor. For birds of prey, commonly used tracking approaches, such as lightweight GPS and inertial measurement units, cannot simultaneously track the bird’s position and the fine details of body part motion or moving external objects. Here we designed a head-mounted, tetherless, 4π camera system which, combined with a high-resolution environmental scan, enabled detailed three-dimensional measurement of both head position and orientation in freely flying Harris’s hawks (Parabuteo unicinctus) performing aerial pursuit in the wild. This stable head-mounted imaging approach enabled reconstruction of flight trajectories in natural environments, detailed quantification of head motion, optic flow and target position, while the bird tracked and captured airborne targets. A head-mounted camera in combination with drone-based mapping of the environment allows reconstruction of the flight trajectory and behavior such as prey capture in freely flying birds such as the Harris’s hawk.
