diff --git a/.dockstore.yml b/.dockstore.yml
index 2a2d4c498b1..2c40ec408fb 100644
--- a/.dockstore.yml
+++ b/.dockstore.yml
@@ -164,7 +164,7 @@ workflows:
          branches:
              - master
              - ah_var_store
-             - gg_VS-1335_BgzipOnExtractAhVarStoreEdition
+             - vs_1351_retry_with_more_memory_vcf_extract
          tags:
              - /.*/
    - name: GvsImportGenomes
@@ -183,7 +183,6 @@ workflows:
          branches:
              - master
              - ah_var_store
-             - rsa_vs_1328
          tags:
              - /.*/
    - name: GvsPrepareRangesCallset
@@ -261,7 +260,6 @@ workflows:
              # `master` is here so there will be *something* under the branches filter, but GVS WDLs are not on `master`
              # so this shouldn't show up in the list of available versions in Dockstore or Terra.
              - master
-             - vs_1099_composite_object_inputs
          tags:
              - /.*/
    - name: GvsJointVariantCallsetCost
@@ -280,6 +278,7 @@ workflows:
          branches:
              - master
              - ah_var_store
+             - rsa_vs_1168_bge
          tags:
              - /.*/
    - name: GvsQuickstartVcfIntegration
@@ -289,7 +288,6 @@ workflows:
          branches:
              - master
              - ah_var_store
-             - gg_VS-1335_BgzipOnExtractAhVarStoreEdition
          tags:
              - /.*/
    - name: GvsQuickstartHailIntegration
@@ -299,7 +297,6 @@ workflows:
          branches:
              - master
              - ah_var_store
-             - gg_VS-1335_BgzipOnExtractAhVarStoreEdition
          tags:
              - /.*/
    - name: GvsQuickstartIntegration
@@ -309,8 +306,7 @@ workflows:
          branches:
              - master
              - ah_var_store
-             - vs_1342_maybe_stick_with_gar
-             - gg_VS-1335_BgzipOnExtractAhVarStoreEdition
+             - vs_1334_vcf_max_alt_alleles
          tags:
              - /.*/
    - name: GvsIngestTieout
@@ -358,6 +354,16 @@ workflows:
              - ah_var_store
          tags:
              - /.*/
+   - name: GvsTieoutVcfMaxAltAlleles
+     subclass: WDL
+     primaryDescriptorPath: /scripts/variantstore/wdl/GvsTieoutVcfMaxAltAlleles.wdl
+     filters:
+       branches:
+         - ah_var_store
+         - master
+         - vs_1334_vcf_max_alt_alleles
+       tags:
+         - /.*/
    - name: HailFromWdl
      subclass: WDL
      primaryDescriptorPath: /scripts/variantstore/wdl/HailFromWdl.wdl
@@ -430,7 +436,6 @@ workflows:
          - master
          - ah_var_store
          - EchoCallset
-         - vs_1315_plink_docker
    - name: MergePgenHierarchicalWdl
      subclass: WDL
      primaryDescriptorPath: /scripts/variantstore/wdl/MergePgenHierarchical.wdl
@@ -446,4 +451,3 @@ workflows:
        branches:
          - ah_var_store
          - EchoCallset
-         - vs_1315_plink_docker
diff --git a/build_docker.sh b/build_docker.sh
index 63b9be5a187..c53c4e51230 100755
--- a/build_docker.sh
+++ b/build_docker.sh
@@ -124,9 +124,9 @@ fi
 
 echo "Building image to tag ${REPO_PRJ}:${GITHUB_TAG}..."
 if [ -n "${IS_PUSH}" ]; then
-    docker build -t ${REPO_PRJ}:${GITHUB_TAG} --build-arg RELEASE=${RELEASE}  --squash .
+    docker build -t ${REPO_PRJ}:${GITHUB_TAG} --build-arg RELEASE=${RELEASE}  --squash . --iidfile /tmp/idfile.txt
 else
-    docker build -t ${REPO_PRJ}:${GITHUB_TAG} --build-arg RELEASE=${RELEASE} .
+    docker build -t ${REPO_PRJ}:${GITHUB_TAG} --build-arg RELEASE=${RELEASE} . --iidfile /tmp/idfile.txt
 fi
 # Since we build the docker image with stages, the first build stage for GATK will be leftover in
 # the local docker context after executing the above commands. This step reclaims that space automatically
diff --git a/scripts/variantstore/docs/Build Docker from VM/building_gatk_docker_on_a_vm.md b/scripts/variantstore/docs/Build Docker from VM/building_gatk_docker_on_a_vm.md
index 6100f17bb82..7f172134e9f 100644
--- a/scripts/variantstore/docs/Build Docker from VM/building_gatk_docker_on_a_vm.md	
+++ b/scripts/variantstore/docs/Build Docker from VM/building_gatk_docker_on_a_vm.md	
@@ -47,16 +47,29 @@ cd gatk
 # Log in to Google Cloud
 gcloud init
 
-# Configure the credential helper for GCR
-gcloud auth configure-docker
+FULL_IMAGE_ID=$(cat /tmp/idfile.txt)
+
+# Take the slice of this full Docker image ID that corresponds with the output of `docker images`:
+IMAGE_ID=${FULL_IMAGE_ID:7:12}
+
+# The GATK Docker image is based on gatkbase.
+IMAGE_TYPE="gatkbase"
+TAG=$(python3 ./scripts/variantstore/wdl/extract/build_docker_tag.py --image-id "${IMAGE_ID}" --image-type "${IMAGE_TYPE}")
+
+BASE_REPO="broad-dsde-methods/gvs"
+REPO_WITH_TAG="${BASE_REPO}/gatk:${TAG}"
+docker tag "${IMAGE_ID}" "${REPO_WITH_TAG}"
+
+# Configure the credential helper for GAR
+gcloud auth configure-docker us-central1-docker.pkg.dev
 
 # Tag and push
-today="$(date -Idate | sed 's/-/_/g')"
-git_hash="$(git rev-parse --short HEAD)"
-DOCKER_TAG="us.gcr.io/broad-dsde-methods/broad-gatk-snapshots:varstore_${today}_${git_hash}"
-DOCKER_IMAGE_ID=$(docker images | head -n 2 | tail -n 1 | awk '{print $3}')
-docker tag $DOCKER_IMAGE_ID $DOCKER_TAG
-docker push $DOCKER_TAG
+GAR_TAG="us-central1-docker.pkg.dev/${REPO_WITH_TAG}"
+docker tag "${REPO_WITH_TAG}" "${GAR_TAG}"
+
+docker push "${GAR_TAG}"
+
+echo "Docker image pushed to \"${GAR_TAG}\""
 ```
 
 Don't forget to shut down (and possibly delete) your VM once you're done!
diff --git a/scripts/variantstore/docs/aou/AOU_DELIVERABLES.md b/scripts/variantstore/docs/aou/AOU_DELIVERABLES.md
index 8ae4b8cf1cd..4636408944a 100644
--- a/scripts/variantstore/docs/aou/AOU_DELIVERABLES.md
+++ b/scripts/variantstore/docs/aou/AOU_DELIVERABLES.md
@@ -83,17 +83,19 @@
     - This workflow needs to be run with the `extract_table_prefix` input from `GvsPrepareRangesCallset` step.
     - This workflow needs to be run with the `filter_set_name` input from `GvsCreateFilterSet` step.
     - This workflow does not use the Terra Data Entity Model to run, so be sure to select the `Run workflow with inputs defined by file paths` workflow submission option.
-1. `GvsExtractCallset` / `GvsExtractCallsetPgenMerged` workflow
+1. `GvsExtractCallset` / `GvsExtractCallsetPgenMerged` workflows ("small callset" Exome, Clinvar, and ACAF threshold extracts in VCF and PGEN formats respectively)
     - You will need to run the `GvsPrepareRangesCallset` workflow for each "[Region](https://support.researchallofus.org/hc/en-us/articles/14929793660948-Smaller-Callsets-for-Analyzing-Short-Read-WGS-SNP-Indel-Data-with-Hail-MT-VCF-and-PLINK)" (interval list) for which a PGEN or VCF deliverable is required for the callset.
         - This workflow transforms the data in the vet, ref_ranges, and samples tables into a schema optimized for extract.
         - The `enable_extract_table_ttl` input should be set to `true` (the default value is `false`), which will add a TTL of two weeks to the tables it creates.
         - `extract_table_prefix` should be set to a name that is unique to the given Region / interval list. See the [naming conventions doc](https://docs.google.com/document/d/1pNtuv7uDoiOFPbwe4zx5sAGH7MyxwKqXkyrpNmBxeow) for guidance on what to use.
         - Specify the `interval_list` appropriate for the PGEN / VCF extraction run you are performing.
+        - `GvsExtractCallsetPgen` currently defaults to 100 alt alleles maximum, which means that any sites having more than that number of alt alleles will be dropped.
         - This workflow does not use the Terra Data Entity Model to run, so be sure to select the `Run workflow with inputs defined by file paths` workflow submission option.
     - Specify the same `call_set_identifier`, `dataset_name`, `project_id`, `extract_table_prefix`, and `interval_list` that were used in the `GvsPrepareRangesCallset` run documented above.
     - Specify the `interval_weights_bed` appropriate for the PGEN / VCF extraction run you are performing. `gs://gvs_quickstart_storage/weights/gvs_full_vet_weights_1kb_padded_orig.bed` is the interval weights BED used for Quickstart.
-    - For `GvsExtractCallset` only, if you want to output VCFs that are compressed using bgzip, set the `bgzip_output_vcfs` input to `true` to generate VCFs that are compressed using bgzip.
-    - For `GvsExtractCallsetPgenMerged` only, select the workflow option "Retry with more memory" and choose a "Memory retry factor" of 1.5
+    - For both `GvsExtractCallset` and `GvsExtractCallsetPgenMerged`, select the workflow option "Retry with more memory" and choose a "Memory retry factor" of 1.5
+    - For `GvsExtractCallset`, make sure to specify the appropriate `maximum_alternate_alleles` value (currently 100).
+    - For `GvsExtractCallset`, if you want to output VCFs that are compressed using bgzip, set the `bgzip_output_vcfs` input to `true` to generate VCFs that are compressed using bgzip.
     - These workflows do not use the Terra Data Entity Model to run, so be sure to select the `Run workflow with inputs defined by file paths` workflow submission option.
 1. `GvsCalculatePrecisionAndSensitivity` workflow
     - Please see the detailed instructions for running the Precision and Sensitivity workflow [here](../../tieout/AoU_PRECISION_SENSITIVITY.md).
diff --git a/scripts/variantstore/wdl/GvsCalculatePrecisionAndSensitivity.wdl b/scripts/variantstore/wdl/GvsCalculatePrecisionAndSensitivity.wdl
index b44e49f0312..38ca31a05a9 100644
--- a/scripts/variantstore/wdl/GvsCalculatePrecisionAndSensitivity.wdl
+++ b/scripts/variantstore/wdl/GvsCalculatePrecisionAndSensitivity.wdl
@@ -10,6 +10,8 @@ workflow GvsCalculatePrecisionAndSensitivity {
     String dataset_name
     String filter_set_name
     File interval_list
+    File? vcf_eval_bed_file
+    Array[String] chromosomes = ["chr20"]
     String project_id
     Array[String] sample_names
 
@@ -32,6 +34,7 @@ workflow GvsCalculatePrecisionAndSensitivity {
 
   parameter_meta {
     call_set_identifier: "The name of the callset for which we are calculating precision and sensitivity."
+    chromosomes: "The chromosome(s) on which to analyze precision and sensitivity. The default value for this is `['chr20']`."
     dataset_name: "The GVS BigQuery dataset name."
     filter_set_name: "The filter_set_name used to generate the callset."
     interval_list: "The intervals over which to calculate precision and sensitivity."
@@ -41,6 +44,7 @@ workflow GvsCalculatePrecisionAndSensitivity {
     truth_vcf_indices: "A list of the VCF indices for the truth data VCFs supplied above."
     truth_beds: "A list of the bed files for the truth data used for analyzing the samples in `sample_names`."
     ref_fasta: "The cloud path for the reference fasta sequence."
+    vcf_eval_bed_file: "Optional bed file for EvaluateVcf; if passed, will be used instead of chromosomes."
   }
 
   String output_basename = call_set_identifier + "_PS"
@@ -132,7 +136,8 @@ workflow GvsCalculatePrecisionAndSensitivity {
         truth_vcf = truth_vcfs[i],
         truth_vcf_index = truth_vcf_indices[i],
         truth_bed = truth_beds[i],
-        interval_list = interval_list,
+        vcf_eval_bed_file = vcf_eval_bed_file,
+        chromosomes = chromosomes,
         output_basename = sample_name + "-bq_roc_filtered",
         is_vqsr_lite = IsVQSRLite.is_vqsr_lite,
         ref_fasta = ref_fasta,
@@ -146,7 +151,8 @@ workflow GvsCalculatePrecisionAndSensitivity {
         truth_vcf = truth_vcfs[i],
         truth_vcf_index = truth_vcf_indices[i],
         truth_bed = truth_beds[i],
-        interval_list = interval_list,
+        vcf_eval_bed_file = vcf_eval_bed_file,
+        chromosomes = chromosomes,
         all_records = true,
         output_basename = sample_name + "-bq_all",
         is_vqsr_lite = IsVQSRLite.is_vqsr_lite,
@@ -377,9 +383,10 @@ task EvaluateVcf {
     File truth_vcf
     File truth_vcf_index
     File truth_bed
+    File? vcf_eval_bed_file
+    Array[String] chromosomes
 
     Boolean all_records = false
-    File interval_list
 
     File ref_fasta
 
@@ -396,6 +403,14 @@ task EvaluateVcf {
   String max_score_field_tag = if (is_vqsr_lite == true) then 'MAX_CALIBRATION_SENSITIVITY' else 'MAX_AS_VQSLOD'
 
   command <<<
+    chromosomes=( ~{sep=' ' chromosomes} )
+
+    echo "Creating .bed file to control which chromosomes should be evaluated."
+    for i in "${chromosomes[@]}"
+    do
+    echo "$i	0	300000000" >> chromosomes.to.eval.txt
+    done
+
     # Prepend date, time and pwd to xtrace log entries.
     PS4='\D{+%F %T} \w $ '
     set -o errexit -o nounset -o pipefail -o xtrace
@@ -403,7 +418,7 @@ task EvaluateVcf {
     rtg format --output human_REF_SDF ~{ref_fasta}
 
     rtg vcfeval \
-      --bed-regions ~{interval_list} \
+      --bed-regions ~{if defined(vcf_eval_bed_file) then vcf_eval_bed_file else "chromosomes.to.eval.txt"} \
       ~{if all_records then "--all-records" else ""} \
       --roc-subset snp,indel \
       --vcf-score-field=INFO.~{max_score_field_tag} \
diff --git a/scripts/variantstore/wdl/GvsExtractCallset.wdl b/scripts/variantstore/wdl/GvsExtractCallset.wdl
index 75a2b0835d8..501f3807f23 100644
--- a/scripts/variantstore/wdl/GvsExtractCallset.wdl
+++ b/scripts/variantstore/wdl/GvsExtractCallset.wdl
@@ -47,6 +47,8 @@ workflow GvsExtractCallset {
     Float y_bed_weight_scaling = 4
     Boolean is_wgs = true
     Boolean convert_filtered_genotypes_to_nocalls = false
+    Boolean write_cost_to_db = true
+    Int? maximum_alternate_alleles
   }
 
   File reference = "gs://gcp-public-data--broad-references/hg38/v0/Homo_sapiens_assembly38.fasta"
@@ -72,7 +74,6 @@ workflow GvsExtractCallset {
 
   Boolean emit_pls = false
   Boolean emit_ads = true
-  Boolean write_cost_to_db = true
 
   String intervals_file_extension = if (zero_pad_output_vcf_filenames) then '-~{output_file_base_name}.interval_list' else '-scattered.interval_list'
   String vcf_extension = if (bgzip_output_vcfs) then '.vcf.bgz' else '.vcf.gz'
@@ -240,6 +241,7 @@ workflow GvsExtractCallset {
         emit_ads                              = emit_ads,
         convert_filtered_genotypes_to_nocalls = convert_filtered_genotypes_to_nocalls,
         write_cost_to_db                      = write_cost_to_db,
+        maximum_alternate_alleles             = maximum_alternate_alleles,
     }
   }
 
@@ -333,6 +335,7 @@ task ExtractTask {
     Int memory_gib
 
     Int? local_sort_max_records_in_ram = 10000000
+    Int? maximum_alternate_alleles
 
     # for call-caching -- check if DB tables haven't been updated since the last run
     String max_last_modified_timestamp
@@ -370,7 +373,18 @@ task ExtractTask {
         --filter-set-name ~{filter_set_name}'
     fi
 
-    gatk --java-options "-Xmx~{memory_gib - 3}g" \
+    # This tool may get invoked with "Retry with more memory" with a different amount of memory than specified in
+    # the input `memory_gib`, so use the memory-related environment variables rather than the `memory_gib` input.
+    # https://support.terra.bio/hc/en-us/articles/4403215299355-Out-of-Memory-Retry
+    if [[ ${MEM_UNIT} == "GB" ]]
+    then
+        memory_mb=$(python3 -c "from math import floor; print(floor((${MEM_SIZE} - 3) * 1000))")
+    else
+        echo "Unexpected memory unit: ${MEM_UNIT}" 1>&2
+        exit 1
+    fi
+
+    gatk --java-options "-Xmx${memory_mb}m" \
       ExtractCohortToVcf \
         --vet-ranges-extract-fq-table ~{fq_ranges_cohort_vet_extract_table} \
         ~{"--vet-ranges-extract-table-version " + vet_extract_table_version} \
@@ -387,6 +401,7 @@ task ExtractTask {
         ~{true='--emit-ads' false='' emit_ads} \
         ~{true='' false='--use-vqsr-scoring' use_VQSR_lite} \
         ~{true='--convert-filtered-genotypes-to-no-calls' false='' convert_filtered_genotypes_to_nocalls} \
+        ~{'--maximum-alternate-alleles ' + maximum_alternate_alleles} \
         ${FILTERING_ARGS} \
         --dataset-id ~{dataset_name} \
         --call-set-identifier ~{call_set_identifier} \
diff --git a/scripts/variantstore/wdl/GvsExtractCallsetPgen.wdl b/scripts/variantstore/wdl/GvsExtractCallsetPgen.wdl
index 6b6063a1795..7a7fe39ec5e 100644
--- a/scripts/variantstore/wdl/GvsExtractCallsetPgen.wdl
+++ b/scripts/variantstore/wdl/GvsExtractCallsetPgen.wdl
@@ -16,8 +16,10 @@ workflow GvsExtractCallsetPgen {
         # reference (see plink chromosome codes here: https://www.cog-genomics.org/plink/2.0/data#irreg_output)
         # ExtractCohortToPgen currently only supports codes "chrM" and "MT"
         String pgen_chromosome_code = "chrM"
-        # Max number of alt alleles a site can have. If a site exceeds this number, it will not be written (max 254)
-        Int max_alt_alleles = 254
+        # Max number of alt alleles a site can have. If a site exceeds this number, it will not be written (max 254).
+        # PGEN extract is currently only used for AoU "small callsets" which always want 100 max alt alleles, so default
+        # to that value.
+        Int max_alt_alleles = 100
         # If true, does not throw an exception for samples@sites with unsupported ploidy (codes it as missing instead)
         Boolean lenient_ploidy_validation = false
         # If true, preserves phasing in the output PGEN files if phasing is present in the source genotypes
@@ -318,7 +320,7 @@ task PgenExtractTask {
         # ExtractCohortToPgen currently only supports codes "chrM" and "MT"
         String pgen_chromosome_code
         # Max number of alt alleles a site can have. If a site exceeds this number, it will not be written (max 254)
-        Int? max_alt_alleles
+        Int max_alt_alleles
         # If true, does not throw an exception for samples@sites with unsupported ploidy (codes it as missing instead)
         Boolean? lenient_ploidy_validation
         # If true, preserves phasing in the output PGEN files if phasing is present in the source genotypes
diff --git a/scripts/variantstore/wdl/GvsJointVariantCalling.wdl b/scripts/variantstore/wdl/GvsJointVariantCalling.wdl
index f8dc056e30c..dfab2a8f76c 100644
--- a/scripts/variantstore/wdl/GvsJointVariantCalling.wdl
+++ b/scripts/variantstore/wdl/GvsJointVariantCalling.wdl
@@ -63,6 +63,8 @@ workflow GvsJointVariantCalling {
 
         File? training_python_script
         File? scoring_python_script
+
+        Int? maximum_alternate_alleles
     }
 
     # If is_wgs is true, we'll use the WGS interval list else, we'll use the Exome interval list.  We'll currently use
@@ -225,6 +227,7 @@ workflow GvsJointVariantCalling {
             drop_state = drop_state,
             bgzip_output_vcfs = bgzip_output_vcfs,
             is_wgs = is_wgs,
+            maximum_alternate_alleles = maximum_alternate_alleles,
     }
 
     output {
diff --git a/scripts/variantstore/wdl/GvsQuickstartHailIntegration.wdl b/scripts/variantstore/wdl/GvsQuickstartHailIntegration.wdl
index d544a6a10f5..91710eae55d 100644
--- a/scripts/variantstore/wdl/GvsQuickstartHailIntegration.wdl
+++ b/scripts/variantstore/wdl/GvsQuickstartHailIntegration.wdl
@@ -35,6 +35,8 @@ workflow GvsQuickstartHailIntegration {
         String? workspace_bucket
         String? workspace_id
         String? submission_id
+
+        Int? maximum_alternate_alleles
     }
 
     String project_id = "gvs-internal"
@@ -89,6 +91,7 @@ workflow GvsQuickstartHailIntegration {
             workspace_bucket = effective_workspace_bucket,
             workspace_id = effective_workspace_id,
             submission_id = effective_submission_id,
+            maximum_alternate_alleles = maximum_alternate_alleles,
     }
 
     call ExtractAvroFilesForHail.GvsExtractAvroFilesForHail {
diff --git a/scripts/variantstore/wdl/GvsQuickstartIntegration.wdl b/scripts/variantstore/wdl/GvsQuickstartIntegration.wdl
index 50e7a836ea2..f038ae69b5b 100644
--- a/scripts/variantstore/wdl/GvsQuickstartIntegration.wdl
+++ b/scripts/variantstore/wdl/GvsQuickstartIntegration.wdl
@@ -28,6 +28,7 @@ workflow GvsQuickstartIntegration {
         String? gatk_docker
         String? hail_version
         Boolean chr20_X_Y_only = true
+        Int? maximum_alternate_alleles
     }
 
     File full_wgs_interval_list = "gs://gcp-public-data--broad-references/hg38/v0/wgs_calling_regions.hg38.noCentromeres.noTelomeres.interval_list"
@@ -102,6 +103,7 @@ workflow GvsQuickstartIntegration {
                 workspace_id = GetToolVersions.workspace_id,
                 submission_id = GetToolVersions.submission_id,
                 hail_version = effective_hail_version,
+                maximum_alternate_alleles = maximum_alternate_alleles,
         }
         call QuickstartHailIntegration.GvsQuickstartHailIntegration as GvsQuickstartHailVQSRClassicIntegration {
             input:
@@ -128,6 +130,7 @@ workflow GvsQuickstartIntegration {
                 workspace_id = GetToolVersions.workspace_id,
                 submission_id = GetToolVersions.submission_id,
                 hail_version = effective_hail_version,
+                maximum_alternate_alleles = maximum_alternate_alleles,
         }
 
         if (GvsQuickstartHailVQSRLiteIntegration.used_tighter_gcp_quotas) {
@@ -173,6 +176,7 @@ workflow GvsQuickstartIntegration {
                 workspace_bucket = GetToolVersions.workspace_bucket,
                 workspace_id = GetToolVersions.workspace_id,
                 submission_id = GetToolVersions.submission_id,
+                maximum_alternate_alleles = maximum_alternate_alleles,
         }
         call QuickstartVcfIntegration.GvsQuickstartVcfIntegration as QuickstartVcfVQSRClassicIntegration {
             input:
@@ -199,6 +203,7 @@ workflow GvsQuickstartIntegration {
                 workspace_bucket = GetToolVersions.workspace_bucket,
                 workspace_id = GetToolVersions.workspace_id,
                 submission_id = GetToolVersions.submission_id,
+                maximum_alternate_alleles = maximum_alternate_alleles,
         }
 
         if (QuickstartVcfVQSRClassicIntegration.used_tighter_gcp_quotas) {
@@ -244,6 +249,7 @@ workflow GvsQuickstartIntegration {
                 workspace_bucket = GetToolVersions.workspace_bucket,
                 workspace_id = GetToolVersions.workspace_id,
                 submission_id = GetToolVersions.submission_id,
+                maximum_alternate_alleles = maximum_alternate_alleles,
         }
 
         if (QuickstartVcfExomeIntegration.used_tighter_gcp_quotas) {
@@ -282,6 +288,7 @@ workflow GvsQuickstartIntegration {
                 workspace_bucket = GetToolVersions.workspace_bucket,
                 workspace_id = GetToolVersions.workspace_id,
                 submission_id = GetToolVersions.submission_id,
+                maximum_alternate_alleles = maximum_alternate_alleles,
                 git_branch_or_tag = git_branch_or_tag,
         }
 
diff --git a/scripts/variantstore/wdl/GvsQuickstartVcfIntegration.wdl b/scripts/variantstore/wdl/GvsQuickstartVcfIntegration.wdl
index e56bdfe3027..570242cfc49 100644
--- a/scripts/variantstore/wdl/GvsQuickstartVcfIntegration.wdl
+++ b/scripts/variantstore/wdl/GvsQuickstartVcfIntegration.wdl
@@ -33,6 +33,8 @@ workflow GvsQuickstartVcfIntegration {
         String? workspace_bucket
         String? workspace_id
         String? submission_id
+
+        Int? maximum_alternate_alleles
     }
     String project_id = "gvs-internal"
 
@@ -110,6 +112,7 @@ workflow GvsQuickstartVcfIntegration {
             git_branch_or_tag = git_branch_or_tag,
             git_hash = effective_git_hash,
             tighter_gcp_quotas = false,
+            maximum_alternate_alleles = maximum_alternate_alleles,
     }
 
     # Only assert identical outputs if we did not filter (filtering is not deterministic) OR if we are using VQSR Lite (which is deterministic)
diff --git a/scripts/variantstore/wdl/GvsTieoutPgenToVcf.wdl b/scripts/variantstore/wdl/GvsTieoutPgenToVcf.wdl
index 9150313d25d..7a84adcf27a 100644
--- a/scripts/variantstore/wdl/GvsTieoutPgenToVcf.wdl
+++ b/scripts/variantstore/wdl/GvsTieoutPgenToVcf.wdl
@@ -55,7 +55,7 @@ task Tieout {
         # Now just the contigs we care about
         zgrep -E '^##' 0000000000-quickit.vcf.gz | grep -E '^##contig=<ID=chr[12]?[0-9],|^##contig=<ID=chr[XY],' >> vcf_header.txt
 
-        # Tab-separated comparision-friendly columns
+        # Tab-separated comparison-friendly columns
         acc=""
         for item in \\#CHROM POS ID REF ALT QUAL FILTER INFO FORMAT $(cat ../samples.txt)
         do
diff --git a/scripts/variantstore/wdl/GvsTieoutVcfMaxAltAlleles.wdl b/scripts/variantstore/wdl/GvsTieoutVcfMaxAltAlleles.wdl
new file mode 100644
index 00000000000..532ec3fd6ab
--- /dev/null
+++ b/scripts/variantstore/wdl/GvsTieoutVcfMaxAltAlleles.wdl
@@ -0,0 +1,98 @@
+version 1.0
+
+import "GvsUtils.wdl" as Utils
+
+workflow GvsTieoutVcfMaxAltAlleles {
+    input {
+        String unfiltered_output_gcs_path
+        String filtered_output_gcs_path
+        Int max_alt_alleles
+        String? variants_docker
+    }
+
+    if (!defined(variants_docker)) {
+        call Utils.GetToolVersions
+    }
+
+    String effective_variants_docker = select_first([variants_docker, GetToolVersions.variants_docker])
+
+    call Tieout {
+        input:
+            unfiltered_output_gcs_path = unfiltered_output_gcs_path,
+            filtered_output_gcs_path = filtered_output_gcs_path,
+            max_alt_alleles = max_alt_alleles,
+            variants_docker = effective_variants_docker,
+    }
+}
+
+task Tieout {
+    input {
+        String unfiltered_output_gcs_path
+        String filtered_output_gcs_path
+        Int max_alt_alleles
+        String variants_docker
+    }
+    command <<<
+        # Prepend date, time and pwd to xtrace log entries.
+        PS4='\D{+%F %T} \w $ '
+        set -o errexit -o nounset -o pipefail -o xtrace
+
+        mkdir unfiltered filtered compare
+
+        cd unfiltered
+        gcloud storage cp '~{unfiltered_output_gcs_path}/*' .
+
+        cd ..
+        cd filtered
+        gcloud storage cp '~{filtered_output_gcs_path}/*' .
+
+        cd ..
+
+        for unfiltered in unfiltered/*.vcf.gz
+        do
+          base="$(basename $unfiltered)"
+          filtered="filtered/${base}"
+          out=${base%-*}
+          bcftools isec $filtered $unfiltered -p compare/${out}
+        done
+
+        # Identify the 'isec' output files that represent entries which are "private" to either VCF file being compared.
+        find . -name README.txt | xargs grep -h private | awk '{print $1}' > private_files.txt
+
+        # Turn off errexit, the grep below is expected to return non-zero
+        set +o errexit
+
+        # Find all the private files, print out all the ALTs, look for any lines that do *not* have at least two commas.
+        # All ALT entries are expected to have two or more commas because the maximum number of ALT alleles was specified
+        # as two. So the unfiltered file would have entries like ALT1,ALT2,...,ALTN while the filtered file would not,
+        # and these should be the only differences between the two files.
+        cat private_files.txt | xargs -n 1 bcftools query -f '%ALT\n' | grep -E -v ',.*,' > unexpected.txt
+
+        rc=$?
+        if [[ $rc -eq 0 ]]
+        then
+            echo "Unexpected ALTs found, see 'unexpected' output for details." 1>&2
+            exit 1
+        fi
+
+        set -o errexit
+
+        # Similar to above, except making sure that we did observe some filtered sites. At the time of this
+        # writing there are 4001 sites filtered for > 2 entries; conservatively setting the threshold at 1000.
+        filtered_count=$(cat private_files.txt | xargs -n 1 bcftools query -f '%ALT\n' | grep -E ',.*,' | wc -l)
+        if [[ $filtered_count -lt 1000 ]]
+        then
+            echo "Unexpectedly found fewer than 1000 filtered ALTs, see 'tarball' output for details." 1>&2
+            exit 1
+        fi
+
+        tar cfvz tieout.tgz compare
+    >>>
+    output {
+        File tarball = "tieout.tgz"
+        File unexpected = "unexpected.txt"
+    }
+    runtime {
+        docker: variants_docker
+    }
+}
diff --git a/scripts/variantstore/wdl/GvsUtils.wdl b/scripts/variantstore/wdl/GvsUtils.wdl
index 40f0fee9651..bfb1b47aff7 100644
--- a/scripts/variantstore/wdl/GvsUtils.wdl
+++ b/scripts/variantstore/wdl/GvsUtils.wdl
@@ -73,8 +73,8 @@ task GetToolVersions {
     # there are a handlful of tasks that require the larger GNU libc-based `slim`.
     String cloud_sdk_slim_docker = "gcr.io/google.com/cloudsdktool/cloud-sdk:435.0.0-slim"
     String variants_docker = "us-central1-docker.pkg.dev/broad-dsde-methods/gvs/variants:2024-04-22-alpine-32804b134a75"
-    String gatk_docker = "us.gcr.io/broad-dsde-methods/broad-gatk-snapshots:varstore_2024_04_24_61922a303"
     String variants_nirvana_docker = "us.gcr.io/broad-dsde-methods/variantstore:nirvana_2022_10_19"
+    String gatk_docker = "us-central1-docker.pkg.dev/broad-dsde-methods/gvs/gatk:2024-05-01-gatkbase-617d4d1c7f64"
     String real_time_genomics_docker = "docker.io/realtimegenomics/rtg-tools:latest"
     String gotc_imputation_docker = "us.gcr.io/broad-gotc-prod/imputation-bcf-vcf:1.0.5-1.10.2-0.1.16-1649948623"
     String plink_docker = "us-central1-docker.pkg.dev/broad-dsde-methods/gvs/plink2:2024-04-23-slim-a0a65f52cc0e"
diff --git a/src/main/java/org/broadinstitute/hellbender/tools/gvs/extract/ExtractCohort.java b/src/main/java/org/broadinstitute/hellbender/tools/gvs/extract/ExtractCohort.java
index 2e230ba44d5..3eae6e1d9aa 100644
--- a/src/main/java/org/broadinstitute/hellbender/tools/gvs/extract/ExtractCohort.java
+++ b/src/main/java/org/broadinstitute/hellbender/tools/gvs/extract/ExtractCohort.java
@@ -145,6 +145,13 @@ public enum VQScoreFilteringType {GENOTYPE, SITES, NONE}
     )
     private boolean convertFilteredGenotypesToNoCalls = false;
 
+    @Argument(
+            fullName = "maximum-alternate-alleles",
+            doc = "The maximum number of alternate alleles a site can have before it is hard-filtered from the output",
+            optional = true
+    )
+    private Long maximumAlternateAlleles = 0L;
+
     @Argument(
             fullName = "snps-truth-sensitivity-filter-level",
             mutex = {"snps-lod-score-cutoff"},
@@ -414,6 +421,10 @@ protected void onStartup() {
                     "if no sample file (--sample-file) is provided.");
         }
 
+        OptionalLong optionalMaximumAlternateAlleles =
+                maximumAlternateAlleles != null && maximumAlternateAlleles > 0L ?
+                        OptionalLong.of(maximumAlternateAlleles) : OptionalLong.empty();
+
         engine = !isVQSR ?
                 new ExtractCohortVETSEngine(
                         projectID,
@@ -441,6 +452,7 @@ protected void onStartup() {
                         emitADs,
                         vqScoreFilteringType,
                         convertFilteredGenotypesToNoCalls,
+                        optionalMaximumAlternateAlleles,
                         inferredReferenceState,
                         presortedAvroFiles,
                         this::apply)
@@ -471,6 +483,7 @@ protected void onStartup() {
                         emitADs,
                         vqScoreFilteringType,
                         convertFilteredGenotypesToNoCalls,
+                        optionalMaximumAlternateAlleles,
                         inferredReferenceState,
                         presortedAvroFiles,
                         this::apply);
diff --git a/src/main/java/org/broadinstitute/hellbender/tools/gvs/extract/ExtractCohortEngine.java b/src/main/java/org/broadinstitute/hellbender/tools/gvs/extract/ExtractCohortEngine.java
index a1f62236e8d..0c201ff1984 100644
--- a/src/main/java/org/broadinstitute/hellbender/tools/gvs/extract/ExtractCohortEngine.java
+++ b/src/main/java/org/broadinstitute/hellbender/tools/gvs/extract/ExtractCohortEngine.java
@@ -46,6 +46,7 @@ public class ExtractCohortEngine {
     private final Double vqScoreINDELThreshold;
     private final ExtractCohort.VQScoreFilteringType vqScoreFilteringType;
     private final boolean convertFilteredGenotypesToNoCalls;
+    private final OptionalLong maxAlternateAlleleCount;
 
     private final String projectID;
     private final boolean emitPLs;
@@ -136,6 +137,7 @@ public ExtractCohortEngine(final String projectID,
                                final boolean emitADs,
                                final ExtractCohort.VQScoreFilteringType vqScoreFilteringType,
                                final boolean convertFilteredGenotypesToNoCalls,
+                               final OptionalLong maxAlternateAlleleCount,
                                final GQStateEnum inferredReferenceState,
                                final boolean presortedAvroFiles,
                                final Consumer<VariantContext> variantContextConsumer
@@ -190,6 +192,7 @@ else if (vetRangesExtractTableVersion == VetRangesExtractVersionEnum.V1) {
         this.vqScoreINDELThreshold = vqScoreINDELThreshold;
         this.vqScoreFilteringType = vqScoreFilteringType;
         this.convertFilteredGenotypesToNoCalls = convertFilteredGenotypesToNoCalls;
+        this.maxAlternateAlleleCount = maxAlternateAlleleCount;
 
         this.filterSetSiteTableRef = vqScoreFilteringType.equals(ExtractCohort.VQScoreFilteringType.NONE) ? null : new TableReference(filterSetSiteTableName, SchemaUtils.FILTER_SET_SITE_FIELDS);
         this.filterSetInfoTableRef = vqScoreFilteringType.equals(ExtractCohort.VQScoreFilteringType.NONE) ? null : new TableReference(filterSetInfoTableName, getFilterSetInfoTableFields());
@@ -583,6 +586,15 @@ VariantContext finalizeCurrentVariant(final List<VariantContext> unmergedVariant
         vcWithRef.genotypes(genotypes);
         mergedVC = vcWithRef.make();
 
+        if (maxAlternateAlleleCount.isPresent()) {
+            if (mergedVC.getAlternateAlleles().size() > maxAlternateAlleleCount.getAsLong()) {
+                    logger.warn("Dropping site on contig {} pos [{}, {}] with {} alt alleles versus maximum {}",
+                            mergedVC.getContig(), mergedVC.getStart(), mergedVC.getEnd(),
+                            mergedVC.getAlternateAlleles().size(), maxAlternateAlleleCount.getAsLong());
+                return null;
+            }
+        }
+
         ReferenceContext referenceContext = new ReferenceContext(refSource, new SimpleInterval(mergedVC));
 
         VariantContext genotypedVC = annotationEngine.annotateContext(mergedVC, new FeatureContext(), referenceContext, null, a -> true);
diff --git a/src/main/java/org/broadinstitute/hellbender/tools/gvs/extract/ExtractCohortVETSEngine.java b/src/main/java/org/broadinstitute/hellbender/tools/gvs/extract/ExtractCohortVETSEngine.java
index 817d9989ffb..87f1ba0c8cd 100644
--- a/src/main/java/org/broadinstitute/hellbender/tools/gvs/extract/ExtractCohortVETSEngine.java
+++ b/src/main/java/org/broadinstitute/hellbender/tools/gvs/extract/ExtractCohortVETSEngine.java
@@ -18,6 +18,7 @@
 import java.util.Map;
 import java.util.Objects;
 import java.util.Optional;
+import java.util.OptionalLong;
 import java.util.function.Consumer;
 import java.util.stream.Stream;
 
@@ -75,6 +76,7 @@ public ExtractCohortVETSEngine(final String projectID,
                                    final boolean emitADs,
                                    final ExtractCohort.VQScoreFilteringType vqScoreFilteringType,
                                    final boolean convertFilteredGenotypesToNoCalls,
+                                   final OptionalLong maximumAlternateAlleles,
                                    final GQStateEnum inferredReferenceState,
                                    final boolean presortedAvroFiles,
                                    final Consumer<VariantContext> variantContextConsumer
@@ -105,6 +107,7 @@ public ExtractCohortVETSEngine(final String projectID,
                 emitADs,
                 vqScoreFilteringType,
                 convertFilteredGenotypesToNoCalls,
+                maximumAlternateAlleles,
                 inferredReferenceState,
                 presortedAvroFiles,
                 variantContextConsumer);