sylvesterkaczmarek · Nov 6, 2026 · Nov 6, 2026 · Nov 7, 2026
diff --git a/docs/api/deprecated.md b/docs/api/deprecated.md
@@ -10,7 +10,6 @@
    :toctree: generated/
 
    pp.subsample
-   pp.recipe_weinreb17
    tl.louvain
    tl.sim
    pl.sim

diff --git a/docs/api/preprocessing.md b/docs/api/preprocessing.md
@@ -8,7 +8,7 @@
 .. currentmodule:: scanpy
 ```
 
-Filtering of highly-variable genes, batch-effect correction, per-cell normalization, preprocessing recipes.
+Filtering of highly-variable genes, batch-effect correction, per-cell normalization.
 
 Any transformation of the data matrix that is not a _tool_. Other than _tools_, preprocessing steps usually don't return an easily interpretable annotation, but perform a basic transformation on the data matrix.
 
@@ -36,16 +36,7 @@ For visual quality control, see {func}`~scanpy.pl.highest_expr_genes` and
    pp.downsample_counts
 ```
 
-## Recipes
-
-```{eval-rst}
-.. autosummary::
-   :nosignatures:
-   :toctree: generated/
-
-   pp.recipe_zheng17
-   pp.recipe_seurat
-```
+For workflows using the removed legacy recipes, see {doc}`../how-to/legacy-recipes`.
 
 (pp-data-integration)=
 

diff --git a/docs/how-to/index.md b/docs/how-to/index.md
@@ -4,6 +4,7 @@ This section contains short examples on how to perform specific tasks with scanp
 
 ```{toctree}
 cell-cycle
+legacy-recipes
 knn-transformers
 plotting-with-marsilea
 ```
diff --git a/docs/how-to/legacy-recipes.md b/docs/how-to/legacy-recipes.md
@@ -0,0 +1,58 @@
+# Migrating from legacy preprocessing recipes
+
+`sc.pp.recipe_seurat`, `sc.pp.recipe_zheng17`, and `sc.pp.recipe_weinreb17` have been removed.
+Compose individual preprocessing functions to make the transformations and their parameters explicit.
+{func}`scanpy.experimental.pp.recipe_pearson_residuals` is unaffected.
+
+The examples below start from non-logarithmized counts and reproduce the final expression matrices of the former Seurat and Zheng recipes with their default arguments and the `ScanpyV1` preset.
+They subset into a new `AnnData` object; start with `adata = adata.copy()` if you also need to preserve the input object.
+Unlike the recipes, calling {func}`~scanpy.pp.log1p` directly on `.X` also records `.uns["log1p"]`.
+Nondefault arguments, especially `log=False`, require separate consideration of the original transformation order.
+
+## Seurat
+
+```python
+import scanpy as sc
+
+sc.pp.filter_cells(adata, min_genes=200)
+sc.pp.filter_genes(adata, min_cells=3)
+sc.pp.normalize_total(adata, target_sum=1e4)
+sc.pp.log1p(adata)
+hvg = sc.pp.highly_variable_genes(
+    adata, flavor="seurat", min_mean=0.0125, max_mean=3, min_disp=0.5, out=True
+)
+adata = adata[:, hvg["highly_variable"]].copy()
+sc.pp.scale(adata, zero_center=True, max_value=10)
+```
+
+## Zheng et al. (2017)
+
+```python
+import scanpy as sc
+
+sc.pp.filter_genes(adata, min_counts=1)
+normalized = sc.pp.normalize_total(adata, out=True)
+adata.X = normalized.normalized
+adata.obs["n_counts_all"] = normalized.norm_factor
+sc.pp.log1p(adata)
+hvg = sc.pp.highly_variable_genes(
+    adata, flavor="cell_ranger", n_top_genes=1000, out=True
+)
+adata = adata[:, hvg["highly_variable"]].copy()
+sc.pp.scale(adata, zero_center=True)
+```
+
+The normalization target is the median nonzero cell total before highly variable gene selection.
+`n_counts_all` retains the former recipe's normalization factors (cell totals divided by that target), rather than raw counts.
+Gene selection uses the log-normalized matrix, matching the former implementation.
+With the default `log=True`, the recipe's second normalization was discarded when it restored its saved log-transformed layer; do not add a second normalization to this workflow.
+To inspect gene dispersion, call {func}`~scanpy.pl.highly_variable_genes` on `hvg` before subsetting (`log=True` for Zheng, `log=False` for the default Seurat recipe).
+
+## Weinreb et al. (2017)
+
+There is no numerically equivalent replacement using only the standard preprocessing functions.
+The former recipe optionally logged **before** normalization, excluded genes exceeding 5% of a cell's total, and scaled by the mean total over all genes.
+It selected genes by mean and coefficient of variation, then computed PCA using population standard deviations plus `0.0001` for standardization.
+In contrast, {func}`~scanpy.pp.scale` uses sample standard deviations, and {func}`~scanpy.pp.highly_variable_genes` uses different selection criteria.
+For new analyses, explicitly choose {func}`~scanpy.pp.normalize_total`, {func}`~scanpy.pp.log1p`, {func}`~scanpy.pp.highly_variable_genes`, {func}`~scanpy.pp.scale`, and {func}`~scanpy.pp.pca` parameters appropriate to the data; this is a change of analysis, not an equivalent substitution.
+To reproduce an existing Weinreb analysis exactly, retain the Scanpy version and parameters used for that analysis.
diff --git a/docs/release-notes/1.13.0b1.md b/docs/release-notes/1.13.0b1.md
@@ -3,7 +3,7 @@
 
 #### Miscellaneous changes
 
-- Deprecate {func}`~scanpy.tl.sim`, {func}`~scanpy.pl.sim`, {func}`~scanpy.datasets.toggleswitch`, {func}`~scanpy.pp.recipe_weinreb17`, {func}`~scanpy.queries.gene_coordinates`, and the undocumented `sc.datasets.burczynski06`, `sc.tl.paga_compare_paths`, `sc.tl.paga_degrees`, `sc.tl.paga_expression_entropies`, and `sc.pl.paga_adjacency`, which have next to no users.
+- Deprecate {func}`~scanpy.tl.sim`, {func}`~scanpy.pl.sim`, {func}`~scanpy.datasets.toggleswitch`, `scanpy.pp.recipe_weinreb17`, {func}`~scanpy.queries.gene_coordinates`, and the undocumented `sc.datasets.burczynski06`, `sc.tl.paga_compare_paths`, `sc.tl.paga_degrees`, `sc.tl.paga_expression_entropies`, and `sc.pl.paga_adjacency`, which have next to no users.
   Remove unreachable code paths from {func}`~scanpy.tl.dpt` {smaller}`L Heumos` ({pr}`4390`)
 
 #### Documentation

diff --git a/docs/release-notes/4493.breaking.md b/docs/release-notes/4493.breaking.md
@@ -0,0 +1,3 @@
+Remove `scanpy.pp.recipe_seurat`, `scanpy.pp.recipe_zheng17`, and `scanpy.pp.recipe_weinreb17` and their private implementation.
+Use explicit preprocessing steps instead; see {doc}`/how-to/legacy-recipes` for migration examples and the numerical differences to consider for Weinreb workflows.
+{func}`scanpy.experimental.pp.recipe_pearson_residuals` remains available.
diff --git a/docs/tutorials/trajectories/paga-paul15.ipynb b/docs/tutorials/trajectories/paga-paul15.ipynb
@@ -196,31 +196,23 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "Apply a simple preprocessing recipe: {func}`scanpy.pp.recipe_zheng17`."
+    "Filter genes, normalize counts, select highly variable genes, and scale explicitly. This preserves the final expression matrix of the former default Zheng recipe: gene selection uses log-normalized values, with no renormalization after selection. See [the migration guide](../../how-to/legacy-recipes.md) for details."
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": null,
    "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "running recipe zheng17\n",
-      "normalizing counts per cell\n",
-      "    finished (0:00:00)\n",
-      "extracting highly variable genes\n",
-      "    finished (0:00:00)\n",
-      "normalizing counts per cell\n",
-      "    finished (0:00:00)\n",
-      "    finished (0:00:00)\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
-    "sc.pp.recipe_zheng17(adata)"
+    "sc.pp.filter_genes(adata, min_counts=1)\n",
+    "sc.pp.normalize_total(adata, out={\"normalized\": \"X\", \"norm_factor\": \"obs.n_counts_all\"})\n",
+    "sc.pp.log1p(adata)\n",
+    "hvg = sc.pp.highly_variable_genes(\n",
+    "    adata, flavor=\"cell_ranger\", n_top_genes=1000, out=True\n",
+    ")\n",
+    "adata = adata[:, hvg[\"highly_variable\"]].copy()\n",
+    "sc.pp.scale(adata, zero_center=True)"
    ]
   },
   {

diff --git a/src/scanpy/external/pp/_scanorama_integrate.py b/src/scanpy/external/pp/_scanorama_integrate.py
@@ -90,7 +90,16 @@ def scanorama_integrate(
     >>> import scanpy as sc
     >>> import scanpy.external as sce
     >>> adata = sc.datasets.pbmc3k()
-    >>> sc.pp.recipe_zheng17(adata)
+    >>> sc.pp.filter_genes(adata, min_counts=1)
+    >>> sc.pp.normalize_total(
+    ...     adata, out={"normalized": "X", "norm_factor": "obs.n_counts_all"}
+    ... )
+    >>> sc.pp.log1p(adata)
+    >>> hvg = sc.pp.highly_variable_genes(
+    ...     adata, flavor="cell_ranger", n_top_genes=1000, out=True
+    ... )
+    >>> adata = adata[:, hvg["highly_variable"]].copy()
+    >>> sc.pp.scale(adata, zero_center=True)
     >>> sc.pp.pca(adata)
 
     We now arbitrarily assign a batch metadata variable to each cell

diff --git a/src/scanpy/preprocessing/__init__.py b/src/scanpy/preprocessing/__init__.py
@@ -18,7 +18,6 @@
 )
 from ._pca import PcaOut, PcaResult, pca
 from ._qc import QcMetricsOut, QcMetricsResult, calculate_qc_metrics
-from ._recipes import recipe_seurat, recipe_weinreb17, recipe_zheng17
 from ._scale import ScaleOut, ScaleResult, scale
 from ._scrublet import ScrubletOut, ScrubletResult, scrublet, scrublet_simulate_doublets
 from ._simple import (
@@ -60,9 +59,6 @@
     "normalize_shifted_clr",
     "normalize_total",
     "pca",
-    "recipe_seurat",
-    "recipe_weinreb17",
-    "recipe_zheng17",
     "regress_out",
     "sample",
     "scale",

diff --git a/src/scanpy/preprocessing/_recipes/__init__.py b/src/scanpy/preprocessing/_recipes/__init__.py