diff --git a/codes/qc_stat_umi.py b/codes/qc_stat_umi.py
new file mode 100755
index 0000000..daaeee7
--- /dev/null
+++ b/codes/qc_stat_umi.py
@@ -0,0 +1,103 @@
+#!/usr/bin/env python3
+# -*-coding:utf-8-*-
+import json
+import os
+import re
+import sys
+
+import numpy as np
+import pandas as pd
+
+if len(sys.argv) != 6:
+    print("usage:python3 qc.py fastp_json_file bamdst_out sort_bamdst_out genecore_json output ")
+    sys.exit()
+
+# out_dir=sys.argv[1]
+# name=sys.argv[2]
+
+fastp_json = sys.argv[1]
+rmdup_bamdst_out = sys.argv[2]
+sort_bamdst_out = sys.argv[3]
+
+genecore_json = sys.argv[4]
+
+out = open(sys.argv[5], 'w')
+
+# fastp_json=''.join((out_dir,'/',name,'.json'))
+# rmdup_json=''.join((out_dir,'/',name,'_rmdup.json'))
+# sorted_cov=''.join((out_dir,'/',name,'_sorted_bamdst/coverage.report'))
+# rmdup_cov=''.join((out_dir,'/',name,'_rmdup_bamdst/coverage.report'))
+# rmdup_depth=''.join((out_dir,'/',name,'_rmdup_bamdst/depth_distribution.plot'))
+# depth_tsv=''.join((out_dir,'/',name,'_rmdup_bamdst/depth.tsv.gz'))
+# rmdup_insert=''.join((out_dir,'/',name,'_rmdup_bamdst/insertsize.plot'))
+#
+# out=open(''.join([out_dir,"/",name,"_qc.txt"]),'w')
+
+with open(fastp_json, 'r') as file:
+    fastp = json.load(file)
+
+with open(os.path.join(sort_bamdst_out, 'coverage.report'), 'r') as file2:
+    bamdst = {}
+    for line in file2:
+        if not re.match(r'^#', line):
+            lines = line.strip().split('\t')
+            bamdst[lines[0]] = lines[1]
+
+with open(os.path.join(rmdup_bamdst_out, 'coverage.report'), 'r') as file3:
+    rmdup_bamdst = {}
+    for line in file3:
+        if not re.match(r'^#', line):
+            lines = line.strip().split('\t')
+            rmdup_bamdst[lines[0]] = lines[1]
+
+with open(genecore_json, 'r') as file4:
+    rmdup_js = json.load(file4)
+
+out.write(''.join(["raw_reads\t", str(fastp["summary"]["before_filtering"]["total_reads"]), "\n"]))
+out.write(''.join(["raw_bases\t", str(fastp["summary"]["before_filtering"]["total_bases"]), "\n"]))
+out.write(''.join(["clean_reads\t", str(fastp["summary"]["after_filtering"]["total_reads"]), "\n"]))
+out.write(''.join(["clean_bases\t", str(fastp["summary"]["after_filtering"]["total_bases"]), "\n"]))
+out.write(''.join(["clean_reads_rate(%)\t", str('%.2f' % (
+            fastp["summary"]["after_filtering"]["total_reads"] / fastp["summary"]["before_filtering"][
+        "total_reads"] * 100)), "\n"]))
+out.write(''.join(["Q20(%)\t", str('%.2f' % (fastp["summary"]["before_filtering"]["q20_rate"] * 100)), "\n"]))
+out.write(''.join(["Q30(%)\t", str('%.2f' % (fastp["summary"]["before_filtering"]["q30_rate"] * 100)), "\n"]))
+out.write(''.join(["mapped_reads\t", str(bamdst["[Total] Mapped Reads"]), "\n"]))
+out.write(''.join(["mapped_rate(%)\t", str('%.2f' % (
+            int(bamdst["[Total] Mapped Reads"]) / fastp["summary"]["after_filtering"]["total_reads"] * 100)), "\n"]))
+#    out.write(''.join(["dup_reads\t",str(bamdst["[Total] PCR duplicate reads"]),"\n"]))
+out.write(''.join(["dup_rate(%)\t", str('%.2f' % (rmdup_js["summary"]["duplication_rate"] * 100)), "\n"]))
+
+out.write(''.join(["probe_bed_size\t", str(bamdst["[Target] Len of region"]), "\n"]))
+out.write(''.join(["target_reads\t", str(bamdst["[Target] Target Reads"]), "\n"]))
+out.write(''.join(
+    ["capture_rate(reads)\t", str(bamdst["[Target] Fraction of Target Reads in mapped reads"].replace('%', '')), "\n"]))
+out.write(''.join(["mean_depth_raw\t", str(bamdst["[Target] Average depth"]), "\n"]))
+out.write(''.join(["mean_depth(dedup)\t", str(rmdup_bamdst["[Target] Average depth(rmdup)"]), "\n"]))
+out.write(''.join(["coverage(>0x)\t", str(rmdup_bamdst["[Target] Coverage (>0x)"].replace('%', '')), "\n"]))
+out.write(''.join(["coverage(>10x)\t", str(rmdup_bamdst["[Target] Coverage (>=10x)"].replace('%', '')), "\n"]))
+out.write(''.join(["coverage(>30x)\t", str(rmdup_bamdst["[Target] Coverage (>=30x)"].replace('%', '')), "\n"]))
+uniformity = round((float(rmdup_bamdst["[Target] Average depth"])) * 0.2)
+
+b = os.popen("sed -n %s'p' %s | cut -f5" % (uniformity, os.path.join(rmdup_bamdst_out, 'depth_distribution.plot')))
+b = float(b.read().strip()) * 100
+out.write(''.join(["coverage(>=0.2*meanx)\t", str('%.2f' % (b)), "\n"]))
+
+depth_tsv_df = pd.read_csv(os.path.join(rmdup_bamdst_out, 'depth.tsv.gz'), compression='gzip', sep="\t")
+coverage_gtoe_80 = round(np.percentile(np.array(depth_tsv_df['Rmdup depth']), 20))
+out.write(''.join(["coverage(>=80%)\t", str(coverage_gtoe_80), "\n"]))
+
+insert_df = pd.read_csv(os.path.join(rmdup_bamdst_out, 'insertsize.plot'), sep="\t", header=None)
+insert_peak = insert_df.loc[insert_df.loc[:, 1].idxmax(), 0]
+insert_avg = ((insert_df.loc[:, 0] * insert_df.loc[:, 1]).sum()) / (insert_df.loc[:, 1].sum())
+out.write(''.join(["insertsize_peak\t", str(insert_peak), "\n"]))
+out.write(''.join(["insertsize_avg\t", str(int(insert_avg)), "\n"]))
+out.write(''.join(["uniq_family_count(ss)\t", str((rmdup_js["summary"]["single_stranded_consensus_sequence"]) + (
+rmdup_js["summary"]["duplex_consensus_sequence"])), "\n"]))
+dup_list = rmdup_js["before_processing"]["duplication_level_histogram"]
+gtoe2 = (sum(dup_list) - dup_list[0]) / sum(dup_list)
+gtoe3 = (sum(dup_list) - dup_list[0] - dup_list[1]) / sum(dup_list)
+out.write(''.join(["family_size_gtoe2(%)\t", str('%.2f' % (gtoe2 * 100)), "\n"]))
+out.write(''.join(["family_size_gtoe3(%)\t", str('%.2f' % (gtoe3 * 100)), "\n"]))
+
+out.close()
diff --git a/wdl/msi.wdl b/wdl/msi.wdl
index f8e15a5..f769668 100755
--- a/wdl/msi.wdl
+++ b/wdl/msi.wdl
@@ -32,7 +32,7 @@ task run_msi {
 
         elif [ "${probe}" == "160" ]; then
 
-        echo "msi 624 探针 "
+        echo "msi 160 探针 "
         msisensor2 msi \
         -M /dataseq/jmdna/software/msisensor2/models_hg19_GRCh37 \
         -t -e $PUBLIC/msi/624_650_160_intersect_74.bed \