Expose additional parameters and document omissions (#168)

cspell.json

@@ -171,6 +171,7 @@
         "Nanopore",
         "PACBIO",
         "unpatterned",
+        "mers",
         // STAR / Picard / GATK option fragments
         "Ftag",
         "Ffile",

tools/CHANGELOG.md

@@ -9,6 +9,8 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/).
 ### Added
 
 - Added WDL implementation for Salmon (`index` and `quant` tasks) [#326](https://github.com/stjudecloud/workflows/pull/326)
+- `kraken2.wdl` `kraken` task now exposes `confidence`, `minimum_hit_groups`, and `quick` parameters [#331](https://github.com/stjudecloud/workflows/pull/331)
+- Documented previously unexposed/undocumented parameters in `star.wdl` (`outSAMtype`, `outMultimapperOrder`) and `qualimap.wdl` (`-p`) using a new `omitted_parameters` `meta` convention [#331](https://github.com/stjudecloud/workflows/pull/331)
 
 ## 2026 August
 

tools/kraken2.wdl

@@ -306,6 +306,12 @@ task build_db {
 task kraken {
     meta {
         description: "Runs Kraken2 on a pair of fastq files"
+        omitted_parameters: [
+            {
+                flag: "--memory-mapping",
+                reason: "Enabling this would reduce RAM usage at the cost of a slower run time and higher disk usage. This trade is usually undesirable. Without further investigation to determine the generalized costs and benefits, this option will not be exposed.",
+            },
+        ]
         outputs: {
             report: {
                 description: "A Kraken2 summary report",
@@ -337,6 +343,12 @@ task kraken {
             group: "Common",
         }
         min_base_quality: "Minimum base quality used in classification"
+        confidence: {
+            description: "Confidence score threshold. Classifications below this threshold are unclassified.",
+            external_help: "https://github.com/DerrickWood/kraken2/blob/master/docs/MANUAL.markdown#confidence-scoring",
+        }
+        minimum_hit_groups: "Minimum number of hit groups (overlapping k-mers sharing the same minimizer) needed to make a classification call"
+        quick: "Stop classification at the first hit instead of an exhaustive k-mer search? Faster but less accurate."
         ncpu: {
             description: "Number of cores to allocate for task",
             group: "Common",
@@ -357,6 +369,9 @@ task kraken {
         Boolean use_names = true
         Boolean use_all_cores = false
         Int min_base_quality = 0
+        Float confidence = 0.0
+        Int minimum_hit_groups = 2
+        Boolean quick = false
         Int ncpu = 4
         Int modify_memory_gb = 0
         Int modify_disk_size_gb = 0
@@ -391,6 +406,9 @@ task kraken {
             --output ~{if store_sequences then "'" + out_sequences + "'" else "-"} \
             --threads "$n_cores" \
             --minimum-base-quality ~{min_base_quality} \
+            --confidence ~{confidence} \
+            --minimum-hit-groups ~{minimum_hit_groups} \
+            ~{if quick then "--quick" else ""} \
             --report "~{out_report}" \
             --report-zero-counts \
             ~{if use_names then "--use-names" else ""} \

tools/qualimap.wdl

@@ -4,7 +4,12 @@ version 1.1
 task rnaseq {
     meta {
         description: "Generates runs QualiMap's rnaseq tool on the input BAM file."
-        help: "Note that we don't expose the `-p` parameter. This is used to set strandedness protocol of the sample, however in practice it only disables certain calculations. We do not expose the parameter so that the full suite of calculations is always performed."
+        omitted_parameters: [
+            {
+                flag: "-p, --sequencing-protocol",
+                reason: "Used to set the strandedness protocol of the sample; in practice it only disables certain calculations. Omitted so that the full suite of calculations is always performed.",
+            },
+        ]
         outputs: {
             raw_summary: "Raw text summary of QualiMap's results. Can be parsed by MultiQC.",
             raw_coverage: "Raw text of QualiMap's coverage analysis results. Can be parsed by MultiQC.",

tools/star.wdl

@@ -146,6 +146,16 @@ task build_star_db {
 task alignment {
     meta {
         description: "Runs the STAR aligner on a set of RNA-Seq FASTQ files"
+        omitted_parameters: [
+            {
+                flag: "--outSAMtype",
+                reason: "STAR's default sort implementation is inefficient in the worst case and can blow up memory/time on some samples. Sorting is disabled here and handled by a dedicated downstream sorting task instead.",
+            },
+            {
+                flag: "--outMultimapperOrder",
+                reason: "Hardcoded to `Random`. The alternative, `Old_2.4`, is a biased quasi-random order used before STAR 2.5.0 and is expected to be deprecated in future STAR releases.",
+            },
+        ]
         external_help: "https://github.com/alexdobin/STAR/blob/2.7.11b/doc/STARmanual.pdf"
         outputs: {
             star_log: {