diff --git a/examples/align_reads_to_reference.py b/examples/align_reads_to_reference.py
new file mode 100644
index 0000000..19616b6
--- /dev/null
+++ b/examples/align_reads_to_reference.py
@@ -0,0 +1,77 @@
+"""Exapmle step pipeline that converts CRAM/BAM file(s) to FASTQ format."""
+
+import os
+import re
+from step_pipeline import pipeline, Backend, Localize, Delocalize, files_exist
+
+
+DOCKER_IMAGE = "weisburd/step-pipeline-examples@sha256:30aed465b319c1d1b42af003f7b37120b84afa1510a6a360de5d6077b67daad6"
+
+REFERENCE_FASTA_PATH = {
+    "37": "gs://gcp-public-data--broad-references/hg19/v0/Homo_sapiens_assembly19.fasta",
+    #"38": "gs://gcp-public-data--broad-references/hg38/v0/Homo_sapiens_assembly38.fasta",
+    "38": "gs://str-truth-set/hg38/ref/hg38.fa",
+    "t2t": "gs://gcp-public-data--broad-references/t2t/v2/chm13v2.0.maskedY.fasta",
+}
+REFERENCE_FASTA_FAI_PATH = {
+    "37": "gs://gcp-public-data--broad-references/hg19/v0/Homo_sapiens_assembly19.fasta.fai",
+    #"38": "gs://gcp-public-data--broad-references/hg38/v0/Homo_sapiens_assembly38.fasta.fai",
+    "38": "gs://str-truth-set/hg38/ref/hg38.fa.fai",
+    "t2t": "gs://gcp-public-data--broad-references/t2t/v2/chm13v2.0.maskedY.fasta.fai",
+}
+
+def main():
+    batch_pipeline = pipeline(backend=Backend.HAIL_BATCH_SERVICE)
+    p = batch_pipeline.get_config_arg_parser()
+    p.add_argument("-o", "--output-cram", required=True, help="gs:// path of output CRAM")
+    p.add_argument("-g", "--reference", choices={"37", "38", "t2t"}, default="38", help="Reference genome fasta.")
+    p.add_argument("fastq_r1", help="Input FASTQ file _R1")
+    p.add_argument("fastq_r2", help="Input FASTQ file _R2")
+    args, _ = p.parse_known_args()
+
+    batch_pipeline.set_name(f"align to {args.reference}")
+
+    if not args.output_cram.endswith(".cram"):
+        p.error("--output-cram path must end with .cram")
+
+    cram_filename = os.path.basename(args.output_cram)
+
+    cpus = 8
+    s1 = batch_pipeline.new_step(f"bwa mem: {cram_filename}", step_number=1, arg_suffix="bwa",
+        image=DOCKER_IMAGE, cpu=cpus, memory="standard", storage="300G")
+    #s1.switch_gcloud_auth_to_user_account()
+
+    local_r1_fastq, local_r2_fastq = s1.inputs(args.fastq_r1, args.fastq_r2, localize_by=Localize.COPY)
+
+    reference_fasta_path = REFERENCE_FASTA_PATH[args.reference]
+    reference_fai_path = REFERENCE_FASTA_FAI_PATH[args.reference]
+
+    local_reference_fasta, _, _, _, _, _, _  = s1.inputs([
+        reference_fasta_path,
+        f"{reference_fasta_path}.fai",
+        f"{reference_fasta_path}.amb",
+        f"{reference_fasta_path}.ann",
+        f"{reference_fasta_path}.bwt",
+        f"{reference_fasta_path}.pac",
+        f"{reference_fasta_path}.sa",
+    ], localize_by=Localize.COPY)
+
+    s1.command("set -ex")
+    s1.command("cd /io")
+    #s1.command(f"bwa index {local_reference_fasta}")
+    s1.command(f"bwa mem -t {cpus} {local_reference_fasta} {local_r1_fastq} {local_r2_fastq} | "
+               f"samtools sort --reference {local_reference_fasta} -o {cram_filename} - ")
+    s1.command(f"samtools index {cram_filename}")
+    #s3.command(f"samtools view -c {prefix}.t2t.cram")
+
+    s1.output(cram_filename, output_path=args.output_cram)
+    s1.output(f"{cram_filename}.crai", output_path=f"{args.output_cram}.crai")
+
+    print(f"Output: {args.output_cram}")
+
+    batch_pipeline.run()
+
+
+if __name__ == "__main__":
+    main()
+
diff --git a/examples/docker/Makefile b/examples/docker/Makefile
index 4cc1420..0cd54d7 100644
--- a/examples/docker/Makefile
+++ b/examples/docker/Makefile
@@ -1,4 +1,5 @@
-TAG = weisburd/step-pipeline-examples:latest
+IMAGE_NAME = step-pipeline-examples
+TAG = weisburd/${IMAGE_NAME}:latest
 
 all: build push
 
@@ -7,3 +8,5 @@ build:
 
 push:
 	docker push $(TAG)
+	docker push $(TAG) | tee /dev/tty | grep sha256 | grep latest | cut -d ' ' -f 3 > sha256.txt
+	cat sha256.txt && [ ! -z "`cat sha256.txt`" ] && sed -i.bak "s/${IMAGE_NAME}@sha256:[^\"]*/${IMAGE_NAME}"@`cat sha256.txt`"/"  ../*.py
\ No newline at end of file