From c52558e4ea1d1a791a41671270cfa913c376a9be Mon Sep 17 00:00:00 2001
From: Robrecht Cannoodt <rcannood@gmail.com>
Date: Tue, 28 Jul 2026 13:53:16 +0200
Subject: [PATCH 1/4] order scdesign2 input by spatial cluster

`simulate_count_scDesign2()` writes its output one cell type at a time, in
contiguous blocks (`new_count[, llim:ulim] <- ...`), but scdesign2 was pulling
the coordinates from the input in its original order. The datasets are not
sorted by spatial_cluster -- MOBNEW starts 1,2,3,3,4,2,4,1 -- so every spot got
the counts of an unrelated spot and the spatial signal was destroyed before
the metrics ever saw it. It shows: scdesign2 came last on morans_I (25.7,
against 1.8 for scdesign3_nb) and on svg_precision.

Every other method already does `input[order(input$obs$spatial_cluster)]`
first; this one now does too.

Also make cell_type_sel and cell_type_prop agree: they are matched by
position, but were built with `unique()` and `table()` respectively, which
only coincide when the labels happen to sort the same way. Both now come from
`unique()` on the sorted data, which also matches the block order for 10 or
more clusters, where numeric and character sorting disagree.
---
 src/methods/scdesign2/script.R | 31 +++++++++++++++++++++----------
 1 file changed, 21 insertions(+), 10 deletions(-)

diff --git a/src/methods/scdesign2/script.R b/src/methods/scdesign2/script.R
index aac648d2..09497d10 100644
--- a/src/methods/scdesign2/script.R
+++ b/src/methods/scdesign2/script.R
@@ -18,13 +18,24 @@ if (par$base != "domain") {
 }
 
 cat("scDesign2 simulation start\n")
-counts <- Matrix::t(input$layers[["counts"]])
-colnames(counts) <- as.character(input$obs$spatial_cluster)
-spatial_cluster_prop <- table(input$obs$spatial_cluster)
+# simulate_count_scDesign2() returns the cells grouped per cell type, so order
+# the spots by spatial cluster first -- otherwise the coordinates we attach
+# below belong to different spots than the counts we attach them to
+ordered_indices <- order(input$obs$spatial_cluster)
+input_ordered <- input[ordered_indices]
+
+counts <- Matrix::t(input_ordered$layers[["counts"]])
+colnames(counts) <- as.character(input_ordered$obs$spatial_cluster)
+
+# cell_type_prop is matched to the fitted models by position, so both have to be
+# in the same order -- and that order has to be the one the spots are sorted in,
+# which unique() gives us since the columns are already grouped per cluster
+cell_types <- unique(colnames(counts))
+spatial_cluster_prop <- prop.table(table(colnames(counts))[cell_types])
 
 copula_result <- scDesign2::fit_model_scDesign2(
   data_mat = counts,
-  cell_type_sel = unique(colnames(counts)),
+  cell_type_sel = cell_types,
   sim_method = "copula",
   ncores = ifelse(!is.null(meta$cpus), meta$cpus, 8L)
 )
@@ -33,11 +44,11 @@ sim_out <- scDesign2::simulate_count_scDesign2(
   copula_result,
   ncol(counts),
   sim_method = "copula",
-  cell_type_prop = prop.table(spatial_cluster_prop)
+  cell_type_prop = spatial_cluster_prop
 )
 
-rownames(sim_out) <- input$var_names
-colnames(sim_out) <- input$obs_names
+rownames(sim_out) <- input_ordered$var_names
+colnames(sim_out) <- input_ordered$obs_names
 
 cat("Generating output\n")
 
@@ -45,10 +56,10 @@ output <- anndataR::AnnData(
   layers = list(
     counts = Matrix::t(sim_out)
   ),
-  obs = input$obs[c("row", "col")],
-  var = input$var,
+  obs = input_ordered$obs[c("row", "col")],
+  var = input_ordered$var,
   uns = c(
-    input$uns,
+    input_ordered$uns,
     list(
       method_id = meta$name
     )

From 3889039481ddbecb37eef7604c6dbe5b8df6860d Mon Sep 17 00:00:00 2001
From: Robrecht Cannoodt <rcannood@gmail.com>
Date: Tue, 28 Jul 2026 14:20:49 +0200
Subject: [PATCH 2/4] update changelog

---
 CHANGELOG.md | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index e3468069..b32f320f 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -1,6 +1,11 @@
 # task_spatial_simulators dev
 
 Bug fixes:
+  - `scdesign2`: order the input by `spatial_cluster` before simulating.
+    `simulate_count_scDesign2()` returns cells grouped per cell type, so the
+    coordinates taken from the unsorted input belonged to different spots than
+    the counts they were attached to. Also make `cell_type_sel` and
+    `cell_type_prop` agree in order, which they did not for 10 or more clusters.
   - `run_benchmark`: raise `uns_length_cutoff` from 15 to 50, so that
     `extract_uns_metadata` no longer drops the `metric_ids` of components that
     emit more than 15 metrics. All 28 `ks_statistic_gene_cell` metrics were

From 7b5a93e1ceb38116d62e77b14cc43eb4a21daac6 Mon Sep 17 00:00:00 2001
From: Robrecht Cannoodt <rcannood@gmail.com>
Date: Tue, 28 Jul 2026 15:00:55 +0200
Subject: [PATCH 3/4] restore the input spot order in scdesign2

Sorting for the simulation is only half of it: the result has to be sorted
back, or scdesign2 joins the group of methods whose output is in a different
spot order than the real dataset and whose clustering_ari is therefore scored
against mismatched spots. See #31.
---
 src/methods/scdesign2/script.R | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/src/methods/scdesign2/script.R b/src/methods/scdesign2/script.R
index 09497d10..ee061fd5 100644
--- a/src/methods/scdesign2/script.R
+++ b/src/methods/scdesign2/script.R
@@ -50,16 +50,21 @@ sim_out <- scDesign2::simulate_count_scDesign2(
 rownames(sim_out) <- input_ordered$var_names
 colnames(sim_out) <- input_ordered$obs_names
 
+# put the spots back in the order they came in. The metrics compare the
+# simulated dataset against the real one spot for spot -- ARI and NMI are
+# invariant to a permutation of the labels, but not of the samples.
+restore_order <- order(ordered_indices)
+
 cat("Generating output\n")
 
 output <- anndataR::AnnData(
   layers = list(
-    counts = Matrix::t(sim_out)
+    counts = Matrix::t(sim_out)[restore_order, , drop = FALSE]
   ),
-  obs = input_ordered$obs[c("row", "col")],
-  var = input_ordered$var,
+  obs = input$obs[c("row", "col")],
+  var = input$var,
   uns = c(
-    input_ordered$uns,
+    input$uns,
     list(
       method_id = meta$name
     )

From 438e84c1b75add1c4c3476535a7b53ee4bc643db Mon Sep 17 00:00:00 2001
From: Robrecht Cannoodt <rcannood@gmail.com>
Date: Tue, 28 Jul 2026 15:00:55 +0200
Subject: [PATCH 4/4] update changelog

---
 CHANGELOG.md | 1 +
 1 file changed, 1 insertion(+)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index b32f320f..37828406 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -6,6 +6,7 @@ Bug fixes:
     coordinates taken from the unsorted input belonged to different spots than
     the counts they were attached to. Also make `cell_type_sel` and
     `cell_type_prop` agree in order, which they did not for 10 or more clusters.
+    The simulated spots are returned in the order they came in.
   - `run_benchmark`: raise `uns_length_cutoff` from 15 to 50, so that
     `extract_uns_metadata` no longer drops the `metric_ids` of components that
     emit more than 15 metrics. All 28 `ks_statistic_gene_cell` metrics were