@@ -4,20 +4,39 @@ History
44
551.5.0 (2026-11-01)
66------------------
7- * Memory usage is reduced through chunking in CCI scoring and permutation
8- testing ~5,000,000 bins @ 20GB in ~20 minutes.
7+ This release contains changes in cell-cell interaction (CCI) mainly
8+ improvements in runtime and memory usage and requires code to be updated.
9+
10+ * Memory usage is reduced through chunking (~256MB) in ```st.cci.run` ``
11+ scoring and permutation testing ~5,000,000 bins @ 20GB in ~20 minutes.
912* SpatialData support added with a new ``st.io `` package. Supporting:
1013 ``atera ``, ``xenium ``, ``visium_hd `` and ``visium ``. Start of
1114 spatial data native plotting with ``st.pl.sdata_plot `` and
1215 ``st.pp.highly_variable_genes ``.
13- * Spatial trajectory and other APIs use cluster labels strings
14- (``cluster="6" ``).
15- * Upgraded ligand-receptor databases to connectomeDB2025 (human and mouse).
16- * Loading via ``load_lrs `` is deprecated for ``lr_pairs `` with
17- ``connectomeDB2020_put `` removed. CCI and GO results change with the
18- new databases.
1916* New datasets available to download via ``visium_hd_sge `` and ``atera_sge ``
2017 and return SpatialData objects.
18+ * Upgraded ligand-receptor databases to connectomeDB2025 (human and mouse).
19+ * Ligand-receptor are an ``LRPairs `` object instead of an array of
20+ 'L_R' strings. It tracks the species and where each pair came from.
21+ Pairs can be filtered and combined (``lrs.filter(...) ``, ``lrs1 | lrs2 ``).
22+ * ``st.tl.cci.run `` takes only ``LRPairs ``; an array of 'L_R' names raises
23+ ``TypeError ``. Wrap strings with
24+ ``st.tl.cci.LRPairs.from_names(names, species="human") ``.
25+ * ``st.tl.cci.lr_pairs(database, species) `` loads a database shipped with
26+ stLearn: ``"connectomeDB2025" `` (default; human or mouse) or
27+ ``"connectomeDB2020" `` (human).
28+ * ``st.tl.cci.read_lr_pairs(path, species=...) ``
29+ reads your own csv/tsv of ligand and receptor columns.
30+ * ``st.tl.cci.load_lrs `` is deprecated; use ``lr_pairs ``, or
31+ ``lr_pairs(...).names() `` for the old array. Note that ``lr_pairs() ``
32+ defaults to connectomeDB2025, where ``load_lrs() `` defaults to
33+ connectomeDB2020.
34+ * ``st.add.lr `` is removed; use ``read_lr_pairs `` or ``LRPairs.from_names ``.
35+ * The LR table from ``run `` stores each pair's source in ``var["source"] ``
36+ and the species in ``uns["lr_species"] ``, which ``run_lr_go `` uses as its
37+ default species.
38+ * Spatial trajectory and other APIs use cluster labels strings
39+ (``cluster="6" ``).
2140* AnnData backwards compatibility available through ``st.io.attach_spatial ``.
2241* CCI results are now a LR table (AnnData) returned by ``st.tl.cci.run ``.
2342 A spots x LR pairs with summaries, scores and p-values in ``var ``.
@@ -28,8 +47,6 @@ History
2847* SME computes weights only between nearby spots and data is kept sparse
2948 where possible.
3049* ``extract_feature `` supports ConvNeXt and takes ``weights= ``.
31- * ``st.tl.cci.run `` takes only ``LRPairs ``; for a list of 'L_R' names use
32- ``st.tl.cci.LRPairs.from_names(names, species="human") ``.
3350* Upgrade libraries: pillow>=13.0.0,<14.0, numba>=0.64.0,
3451 scikit-image>=0.25.0, torch>=2.5.0, matplotlib>=3.11,<4.0.
3552
0 commit comments