https://drieslab.r-universe.devRecent package updates in drieslabR-universehttps://github.com/drieslab.pnghttps://drieslab.r-universe.devWed, 07 Oct 2026 18:27:40 GMTrdries@bu.edu (Ruben Dries)Discovery and statistical evaluation of recurrent multi-cellular motifs (size 2, 3, and 4) in spatial omics data. Provides a lightweight 'SpatialGraph' representation backed by parquet, enumeration over igraph, spatially-aware null models, motif-to-gene association, and cross-region or cross-sample comparison. Designed to be framework-agnostic so that the core can be folded into Giotto Suite later.https://github.com/r-universe/drieslab/actions/runs/37823185357Wed, 07 Oct 2026 18:27:40 GMTsmotif0.2.0successhttps://drieslab.r-universe.devhttps://github.com/drieslab/smotifsmotif.Rmdsmotif.htmlSpatial multi-cellular motifs with smotif2026-05-15 01:15:492026-05-15 01:15:49rdries@bu.edu (Ruben Dries)Drop-in Rust acceleration for the smotif R package: parallel enumeration of size-2 / size-3 / size-4 multi-cellular motifs in spatial omics graphs, plus direct parquet I/O for million-cell datasets. Distributed as pre-compiled binaries via R-universe.https://github.com/r-universe/drieslab/actions/runs/37823176451Wed, 07 Oct 2026 18:27:32 GMTsmotifrs0.2.0failurehttps://drieslab.r-universe.devhttps://github.com/drieslab/smotifrsvjarzabe@bu.edu (Veronica Jarzabek)Streams a gzipped CosMx exprMat CSV and emits nonzero triplets.https://github.com/r-universe/drieslab/actions/runs/38649952196Fri, 18 Sep 2026 18:30:45 GMTcosmxscan0.1.0successhttps://drieslab.r-universe.devhttps://github.com/drieslab/cosmxscanthedrieslab@gmail.com (Jiaji George Chen)An extensible framework for tile-based processing of large, spatially indexed data. Tile plans are first-class objects describing a set of bounds, which may be regular grids in pixel or coordinate space, tiles centered on supplied points, or adaptive decompositions such as quadtrees. Bounds may carry padding, supplying the overlapping context that neighborhood operations and objects spanning several tiles require. Extraction dispatches on the class of the input, so one plan can window two data sources of differing type together; methods are provided for 'terra' rasters and vector proxies, and other packages can register methods for their own formats. Tiles are processed in parallel through 'future', or optionally 'BiocParallel', and a per-worker initialization step amortizes expensive setup such as loading a model.https://github.com/r-universe/drieslab/actions/runs/37823185714Thu, 17 Sep 2026 06:53:43 GMTtilework1.0.0successhttps://drieslab.r-universe.devhttps://github.com/drieslab/tileworktile-plans.Rmdtile-plans.htmlChoosing and Creating a Tile Plan2026-04-06 19:02:332026-04-08 04:15:34patch-feature-extraction.Rmdpatch-feature-extraction.htmlPatch-Based Feature Extraction for Machine Learning2026-04-06 18:05:122026-04-08 04:15:34tile-orchestration.Rmdtile-orchestration.htmlSelecting, Grouping, and Iterating Tiles2026-04-06 19:02:332026-07-31 16:42:07