diff --git a/pipelines/_ingredients/wepDirectDetectScienceGroupPipeline.yaml b/pipelines/_ingredients/wepDirectDetectScienceGroupPipeline.yaml
index c02dc4d8..b470bbf7 100644
--- a/pipelines/_ingredients/wepDirectDetectScienceGroupPipeline.yaml
+++ b/pipelines/_ingredients/wepDirectDetectScienceGroupPipeline.yaml
@@ -26,19 +26,6 @@ tasks:
       estimateZernikes.saveHistory: False
       estimateZernikes.maskKwargs: { "doMaskBlends": False }
       donutStampSelector.maxSelect: 5
-  isr:
-    class: lsst.ip.isr.IsrTaskLSST
-    config:
-      # Although we don't have to apply the amp offset corrections, we do want
-      # to compute them for analyzeAmpOffsetMetadata to report on as metrics.
-      doAmpOffset: true
-      ampOffset.doApplyAmpOffset: false
-      # Turn off slow steps in ISR
-      doBrighterFatter: false
-      # Mask saturated pixels,
-      # but turn off quadratic crosstalk because it's currently broken
-      doSaturation: True
-      crosstalk.doQuadraticCrosstalkCorrection: False
 
 subsets:
   wepDirectDetect:
diff --git a/pipelines/production/comCamDailyProcessing.yaml b/pipelines/production/comCamDailyProcessing.yaml
index 667f33b0..9edf3eae 100644
--- a/pipelines/production/comCamDailyProcessing.yaml
+++ b/pipelines/production/comCamDailyProcessing.yaml
@@ -14,7 +14,6 @@ tasks:
 subsets:
   step1:
     subset:
-      - isr
       - generateDonutDirectDetectTask
       - cutOutDonutsScienceSensorGroupTask
       - calcZernikesTask
diff --git a/pipelines/production/comCamRapidAnalysisDanishPipeline.yaml b/pipelines/production/comCamRapidAnalysisDanishPipeline.yaml
index 4b35f076..676f8117 100644
--- a/pipelines/production/comCamRapidAnalysisDanishPipeline.yaml
+++ b/pipelines/production/comCamRapidAnalysisDanishPipeline.yaml
@@ -5,6 +5,19 @@ imports:
   - $TS_WEP_DIR/pipelines/_ingredients/donutVizGroupPipeline.yaml
 
 tasks:
+  isr:
+    class: lsst.ip.isr.IsrTaskLSST
+    config:
+      # Although we don't have to apply the amp offset corrections, we do want
+      # to compute them for analyzeAmpOffsetMetadata to report on as metrics.
+      doAmpOffset: true
+      ampOffset.doApplyAmpOffset: false
+      # Turn off slow steps in ISR
+      doBrighterFatter: false
+      # Mask saturated pixels,
+      # but turn off quadratic crosstalk because it's currently broken
+      doSaturation: True
+      crosstalk.doQuadraticCrosstalkCorrection: False
   calcZernikesTask:
     class: lsst.ts.wep.task.calcZernikesTask.CalcZernikesTask
     config:
diff --git a/pipelines/production/comCamRapidAnalysisPipeline.yaml b/pipelines/production/comCamRapidAnalysisPipeline.yaml
index b210c3bb..a51a6169 100644
--- a/pipelines/production/comCamRapidAnalysisPipeline.yaml
+++ b/pipelines/production/comCamRapidAnalysisPipeline.yaml
@@ -5,6 +5,19 @@ imports:
   - $TS_WEP_DIR/pipelines/_ingredients/donutVizGroupPipeline.yaml
 
 tasks:
+  isr:
+    class: lsst.ip.isr.IsrTaskLSST
+    config:
+      # Although we don't have to apply the amp offset corrections, we do want
+      # to compute them for analyzeAmpOffsetMetadata to report on as metrics.
+      doAmpOffset: true
+      ampOffset.doApplyAmpOffset: false
+      # Turn off slow steps in ISR
+      doBrighterFatter: false
+      # Mask saturated pixels,
+      # but turn off quadratic crosstalk because it's currently broken
+      doSaturation: True
+      crosstalk.doQuadraticCrosstalkCorrection: False
   calcZernikesTask:
     class: lsst.ts.wep.task.calcZernikesTask.CalcZernikesTask
     config: