[TheiaCoV_Illumina_SE] Fix kraken2 raw input and expose kraken2 dehos…

…ted outputs (#597) * make kraken 2 raw actually run on the raw reads; expose kraken2 dehosted outputs * update md5sums * update md5sum * kraken2 raw is now actually running on the raw * update CI
theiagen · Aug 29, 2024 · 5be3433 · 5be3433
1 parent 26c1115
commit 5be3433
Show file tree

Hide file tree

Showing 4 changed files with 12 additions and 9 deletions.
diff --git a/tests/workflows/theiacov/test_wf_theiacov_illumina_se.yml b/tests/workflows/theiacov/test_wf_theiacov_illumina_se.yml
@@ -74,7 +74,7 @@
     - path: miniwdl_run/call-read_QC_trim/call-fastq_scan_raw/work/VERSION
     # kraken2
     - path: miniwdl_run/call-read_QC_trim/call-kraken2_theiacov_raw/command
-      md5sum: 0efd43fc9f7079a38e5b094245c97f59
+      md5sum: ca22e45a62c5c26c4447cdafe75a26ab
     - path: miniwdl_run/call-read_QC_trim/call-kraken2_theiacov_raw/inputs.json
       contains: ["read1", "samplename"]
     - path: miniwdl_run/call-read_QC_trim/call-kraken2_theiacov_raw/outputs.json
@@ -85,14 +85,13 @@
     - path: miniwdl_run/call-read_QC_trim/call-kraken2_theiacov_raw/task.log
       contains: ["wdl", "theiacov_illumina_se", "kraken2_theiacov_raw", "done"]
     - path: miniwdl_run/call-read_QC_trim/call-kraken2_theiacov_raw/work/PERCENT_HUMAN
-      md5sum: 4fd4dcef994592f9865e9bc8807f32f4
+      md5sum: 1576d5d341223ea9d44b0b8a213bb9da
     - path: miniwdl_run/call-read_QC_trim/call-kraken2_theiacov_raw/work/PERCENT_SC2
-      md5sum: 73f7f6bf2257905cb4bee12d23247db3
+      md5sum: 7cc2eb659e21f15fa902b11812eae1f6
     - path: miniwdl_run/call-read_QC_trim/call-kraken2_theiacov_raw/work/PERCENT_TARGET_ORGANISM
       md5sum: 68b329da9893e34099c7d8ad5cb9c940
-    - path: miniwdl_run/call-read_QC_trim/call-kraken2_theiacov_raw/work/_miniwdl_inputs/0/ERR6319327_1.clean.fastq.gz
     - path: miniwdl_run/call-read_QC_trim/call-kraken2_theiacov_raw/work/ERR6319327_kraken2_report.txt
-      md5sum: 3d1bab1c5040916642df5baf20a2d6bd
+      md5sum: 9a089b8920e55c9cc7bc8cd7d18f9a8e
     # clean read screen
     - path: miniwdl_run/call-clean_check_reads/command
       md5sum: aec6c57452ddff84c325601a780605d2

diff --git a/tests/workflows/theiaprok/test_wf_theiaprok_illumina_se.yml b/tests/workflows/theiaprok/test_wf_theiaprok_illumina_se.yml
@@ -598,6 +598,6 @@
     - path: miniwdl_run/wdl/workflows/utilities/wf_merlin_magic.wdl
       md5sum: 359bc59b671ef3121f6b4ac589f0c80e
     - path: miniwdl_run/wdl/workflows/utilities/wf_read_QC_trim_se.wdl
-      md5sum: 51c696b5d122eb1bc7e03d31ccc82cc6
+      md5sum: d11bfe33fdd96eab28892be5a01c1c7d
     - path: miniwdl_run/workflow.log
       contains: ["wdl", "theiaprok_illumina_se", "NOTICE", "done"]
diff --git a/workflows/theiacov/wf_theiacov_illumina_se.wdl b/workflows/theiacov/wf_theiacov_illumina_se.wdl
@@ -232,12 +232,16 @@ workflow theiacov_illumina_se {
     File? read1_clean = read_QC_trim.read1_clean
     String? bbduk_docker = read_QC_trim.bbduk_docker
     # Read QC - kraken outputs
+    String? kraken_version = read_QC_trim.kraken_version
     Float? kraken_human = read_QC_trim.kraken_human
     Float? kraken_sc2 = read_QC_trim.kraken_sc2
     String? kraken_target_organism = read_QC_trim.kraken_target_organism
     String? kraken_target_organism_name = read_QC_trim.kraken_target_organism_name
-    String? kraken_version = read_QC_trim.kraken_version
     File? kraken_report = read_QC_trim.kraken_report
+    Float? kraken_human_dehosted = read_QC_trim.kraken_human_dehosted
+    Float? kraken_sc2_dehosted = read_QC_trim.kraken_sc2_dehosted
+    String? kraken_target_organism_dehosted = read_QC_trim.kraken_target_organism_dehosted
+    File? kraken_report_dehosted = read_QC_trim.kraken_report_dehosted
     # Read Alignment - bwa outputs
     String? bwa_version = ivar_consensus.bwa_version
     String? samtools_version = ivar_consensus.samtools_version

diff --git a/workflows/utilities/wf_read_QC_trim_se.wdl b/workflows/utilities/wf_read_QC_trim_se.wdl
@@ -99,7 +99,7 @@ workflow read_QC_trim_se {
     call kraken.kraken2_theiacov as kraken2_theiacov_raw {
       input:
         samplename = samplename,
-        read1 = bbduk_se.read1_clean,
+        read1 = read1,
         target_organism = target_organism
     }
     call kraken.kraken2_theiacov as kraken2_theiacov_dehosted {
@@ -158,7 +158,7 @@ workflow read_QC_trim_se {
     File? fastqc_raw1_html = fastqc_raw.read1_fastqc_html
     File? fastqc_clean1_html = fastqc_clean.read1_fastqc_html
 
-    # kraken2
+    # kraken2 - raw and dehosted
     String kraken_version = select_first([kraken2_theiacov_raw.version, kraken2_standalone.kraken2_version, ""])
     Float? kraken_human = kraken2_theiacov_raw.percent_human
     Float? kraken_sc2 = kraken2_theiacov_raw.percent_sc2