diff --git a/README.md b/README.md
index 86fd5697..7cbe8d46 100644
--- a/README.md
+++ b/README.md
@@ -34,12 +34,14 @@ data to biological insights.
 
 ### Filtering and Selection Functions
 
-| Function              | Description                                         |
-|-----------------------|-----------------------------------------------------|
-| `identify_abundant()` | Identify abundant transcripts without removing them |
-| `keep_abundant()`     | Keep abundant transcripts                           |
-| `keep_variable()`     | Keep variable transcripts                           |
-| `filterByExpr()`      | Filter by expression                                |
+| Function                                | Description                                                            |
+|-----------------------------------------|------------------------------------------------------------------------|
+| `identify_abundant()`                   | Identify abundant transcripts without removing them                    |
+| `keep_abundant()`                       | Keep abundant transcripts                                              |
+| `identify_abundant_per_category()`      | Identify abundant transcripts per experimental category                |
+| `keep_abundant_per_category()`          | Keep transcripts abundant in at least one experimental category        |
+| `keep_variable()`                       | Keep variable transcripts                                              |
+| `filterByExpr()`                        | Filter by expression                                                   |
 
 ### Dimensionality Reduction Functions
 
diff --git a/inst/NEWS.rd b/inst/NEWS.rd
index b72a4f6c..c6009c82 100644
--- a/inst/NEWS.rd
+++ b/inst/NEWS.rd
@@ -4,6 +4,7 @@
 \section{Changes in development version}{
 \itemize{
     \item \strong{BREAKING CHANGE:} \code{reduce_dimensions()} no longer accepts \code{.abundance} (a tidy-eval assay symbol). The assay must be given explicitly as a character string via \code{assay = "..."}, matching \code{assayNames(object)}. Omitting \code{assay} or passing an unknown assay name raises an error, so users must deliberately choose which abundance matrix to use (commonly a library-size scaled assay from \code{scale_abundance()}). Update all calls, tests, and vignettes that used \code{.abundance = counts} to \code{assay = "counts"} (or the appropriate assay name).
+    \item Added \code{identify_abundant_per_category()} and \code{keep_abundant_per_category()} as category-aware alternatives to \code{identify_abundant()} and \code{keep_abundant()}. These functions evaluate abundance within each experimental group independently and retain features that are sufficiently expressed in at least one category, making it possible to preserve group-specific or condition-specific signals that would be lost by global abundance filtering. The grouping is specified via \code{formula_design} (e.g. \code{~dex}), and the minimum number of categories required is controlled by \code{minimum_category} (default 1).
 }}
 
 \section{Changes in version 1.2.0, Bioconductor 3.12 Release}{
diff --git a/vignettes/introduction.Rmd b/vignettes/introduction.Rmd
index 92bcdadb..fde0b711 100644
--- a/vignettes/introduction.Rmd
+++ b/vignettes/introduction.Rmd
@@ -47,6 +47,8 @@ Function | Description
 ------------ | -------------
 `identify_abundant()` | Identify abundant transcripts without removing them
 `keep_abundant()` | Keep abundant transcripts
+`identify_abundant_per_category()` | Identify abundant transcripts per experimental category
+`keep_abundant_per_category()` | Keep transcripts abundant in at least one experimental category
 `keep_variable()` | Keep variable transcripts
 `filterByExpr()` | Filter by expression
 
@@ -328,6 +330,95 @@ airway = airway_abundant_formula
 
 > **Tip:** Use `formula_design` for complex designs, and use the CPM threshold for library-size-aware filtering.
 
+### Category-Aware Abundance Filtering
+
+Standard abundance filtering (`keep_abundant()`) asks whether a feature is sufficiently expressed across *all* samples. This works well when the signal of interest is globally abundant. However, it can silently discard features that are strongly expressed in only one experimental group — for example, a gene that is highly induced by treatment but nearly absent in the control condition.
+
+`keep_abundant_per_category()` addresses this by evaluating abundance *within each experimental category independently*. A feature is retained if it satisfies the abundance threshold in **at least one** category, regardless of its expression in the other categories.
+
+This approach is particularly important for:
+
+- **Case/control studies** where disease biomarkers are expressed only in the case group
+- **Treatment-response experiments** where genes are induced by a specific stimulus
+- **Multi-tissue or multi-cell-type datasets** with tissue-specific or cell-type-specific transcripts
+- **Subgroup discovery** analyses where group-specific signals would be lost by global filtering
+
+#### Why this matters
+
+Consider a gene expressed at high levels in the `trt` group but near-zero in `untrt`. Global filtering may remove it because its average across all samples is low. But that gene is exactly what you want to detect in a differential expression test. Category-aware filtering preserves it.
+
+#### Run per-category filtering
+
+```{r filtering-per-category}
+# Keep genes abundant in at least one treatment group
+airway_abundant_per_category = airway_abundant_formula |>
+  keep_abundant_per_category(
+    formula_design = ~dex,       # experimental grouping variable
+    minimum_counts = 10,         # minimum raw count threshold
+    minimum_proportion = 0.7     # fraction of samples in a category that must pass
+  )
+```
+
+#### Using CPM threshold instead of raw counts
+
+```{r filtering-per-category-cpm}
+# Use counts-per-million threshold for library-size-aware filtering
+airway_abundant_per_category_cpm = airway_abundant_formula |>
+  keep_abundant_per_category(
+    formula_design = ~dex,
+    minimum_count_per_million = 1,
+    minimum_proportion = 0.7
+  )
+```
+
+#### Require abundance in multiple categories
+
+By default (`minimum_category = 1`) a feature is kept if it passes the threshold in at least one category. Set `minimum_category = 2` to require abundance in at least two categories:
+
+```{r filtering-per-category-min2}
+airway_abundant_per_category_min2 = airway_abundant_formula |>
+  keep_abundant_per_category(
+    formula_design = ~dex,
+    minimum_counts = 10,
+    minimum_proportion = 0.7,
+    minimum_category = 2   # must be abundant in both categories
+  )
+```
+
+#### Mark features without filtering
+
+If you only want to annotate features with a logical `.abundant` column without immediately filtering, use the companion function `identify_abundant_per_category()`:
+
+```{r filtering-per-category-identify}
+airway_marked = airway_abundant_formula |>
+  identify_abundant_per_category(formula_design = ~dex)
+```
+
+#### Compare global vs. category-aware filtering
+
+```{r filtering-compare-global-vs-category}
+# Summary statistics
+airway_abundant_formula |> summarise(
+  filtering = "global",
+  n_features = n_distinct(.feature)
+)
+
+airway_abundant_per_category |> summarise(
+  filtering = "per category (>=1)",
+  n_features = n_distinct(.feature)
+)
+
+airway_abundant_per_category_min2 |> summarise(
+  filtering = "per category (>=2)",
+  n_features = n_distinct(.feature)
+)
+```
+
+> **When to use category-aware filtering:**
+> Use `keep_abundant_per_category()` when your experimental design contains distinct biological groups and you want to retain features that are informative for at least one of them. Use global `keep_abundant()` when you are interested only in features that are broadly expressed across all samples.
+
+
+
 ### Remove Redundant Transcripts
 
 Redundancy removal is a standard approach for reducing highly correlated features.