Merge branch 'CW-3375' into 'dev'

Drop mapula [CW-3375] Closes CW-3375 See merge request epi2melabs/workflows/wf-human-variation!188
epi2me-labs · Jan 23, 2024 · c4c203d · c4c203d
2 parents 4aed60f + 439bdd2
commit c4c203d
Show file tree

Hide file tree

Showing 6 changed files with 8 additions and 45 deletions.
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
@@ -139,28 +139,28 @@ docker-run:
           variables:
               NF_BEFORE_SCRIPT: "wget -q -O demo_data.tar.gz https://ont-exd-int-s3-euwst1-epi2me-labs.s3.amazonaws.com/wf-human-methyl/demo_data.tar.gz && tar -xzvf demo_data.tar.gz"
               NF_PROCESS_FILES: "workflows/methyl.nf"
-              NF_IGNORE_PROCESSES: "index_ref_gzi,mapula,getGenome,get_coverage,getVersions,getParams,failedQCReport,makeAlignmentReport,annotate_vcf,eval_downsampling,downsampling,compress_bed,modkit,modkit_phase"
+              NF_IGNORE_PROCESSES: "index_ref_gzi,getGenome,get_coverage,getVersions,getParams,failedQCReport,makeAlignmentReport,annotate_vcf,eval_downsampling,downsampling,compress_bed,modkit,modkit_phase"
               NF_WORKFLOW_OPTS: "--snp false --sv false --mod --cnv false --str false --bam demo_data/demo.ucram --ref demo_data/GCA_000001405.15_GRCh38_no_alt_analysis_set.fa.gz --sample_name HOOT --ubam_map_threads 2 --ubam_sort_threads 1 --ubam_bam2fq_threads 1 --bam_min_coverage 0.00001"
 
         - if: $MATRIX_NAME == "gzref-bam-mod-hap"
           variables:
               NF_BEFORE_SCRIPT: "wget -q -O demo_data.tar.gz https://ont-exd-int-s3-euwst1-epi2me-labs.s3.amazonaws.com/wf-human-methyl/demo_data.tar.gz && tar -xzvf demo_data.tar.gz"
               NF_PROCESS_FILES: "workflows/methyl.nf"
-              NF_IGNORE_PROCESSES: "minimap2_ubam,index_ref_gzi,mapula,get_coverage,getVersions,getParams,failedQCReport,makeAlignmentReport,annotate_vcf,eval_downsampling,downsampling,modkit"
+              NF_IGNORE_PROCESSES: "minimap2_ubam,index_ref_gzi,get_coverage,getVersions,getParams,failedQCReport,makeAlignmentReport,annotate_vcf,eval_downsampling,downsampling,modkit"
               NF_WORKFLOW_OPTS: "--snp false --sv false --mod --cnv false --str false --bam demo_data/demo.cram --ref demo_data/GCA_000001405.15_GRCh38_no_alt_analysis_set.fa.gz --sample_name HOOT --ubam_map_threads 2 --ubam_sort_threads 1 --ubam_bam2fq_threads 1 --bam_min_coverage 0.00001 --phased --annotation false"
 
         - if: $MATRIX_NAME == "gzref-bam-mod"
           variables:
               NF_BEFORE_SCRIPT: "wget -q -O demo_data.tar.gz https://ont-exd-int-s3-euwst1-epi2me-labs.s3.amazonaws.com/wf-human-methyl/demo_data.tar.gz && tar -xzvf demo_data.tar.gz"
               NF_PROCESS_FILES: "workflows/methyl.nf"
-              NF_IGNORE_PROCESSES: "minimap2_ubam,index_ref_gzi,mapula,get_coverage,getVersions,getParams,failedQCReport,makeAlignmentReport,annotate_vcf,eval_downsampling,downsampling,modkit_phase"
+              NF_IGNORE_PROCESSES: "minimap2_ubam,index_ref_gzi,get_coverage,getVersions,getParams,failedQCReport,makeAlignmentReport,annotate_vcf,eval_downsampling,downsampling,modkit_phase"
               NF_WORKFLOW_OPTS: "--snp false --sv false --mod --cnv false --str false --bam demo_data/demo.cram --ref demo_data/GCA_000001405.15_GRCh38_no_alt_analysis_set.fa.gz --sample_name HOOT --ubam_map_threads 2 --ubam_sort_threads 1 --ubam_bam2fq_threads 1 --bam_min_coverage 0.00001 --annotation false"
 
         - if: $MATRIX_NAME == "downsample"
           variables:
               NF_BEFORE_SCRIPT: "wget -qO demo_data.tar.gz https://ont-exd-int-s3-euwst1-epi2me-labs.s3.amazonaws.com/wf-human-str/demo_data.tar.gz && tar -xzvf demo_data.tar.gz"
               NF_PROCESS_FILES: "workflows/methyl.nf"
-              NF_IGNORE_PROCESSES: "minimap2_ubam,index_ref_gzi,mapula,getGenome,get_coverage,getVersions,getParams,failedQCReport,makeAlignmentReport,eval_downsampling,downsampling,compress_bed,modkit_phase,annotate_vcf"
+              NF_IGNORE_PROCESSES: "minimap2_ubam,index_ref_gzi,getGenome,get_coverage,getVersions,getParams,failedQCReport,makeAlignmentReport,eval_downsampling,downsampling,compress_bed,modkit_phase,annotate_vcf"
               NF_WORKFLOW_OPTS: "--snp false --sv false --mod --annotation false --bam demo_data/demo.bam --ref demo_data/GCA_000001405.15_GRCh38_no_alt_analysis_set.fa.gz \
               --sample_name HOOT --ubam_map_threads 2 --ubam_sort_threads 1 \
               --ubam_bam2fq_threads 1 --bam_min_coverage 0 --skip_annotation \

diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -26,6 +26,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - wf-basecalling subworkflow
     - fast5_dir input and other basecalling related options have been removed from the workflow parameters
     - Users should run the standalone wf-basecalling workflow and provide the output to wf-human-variation
+- Mapula statistics with `--mapula`
 
 ## [v1.9.2]
 ### Fixed

diff --git a/main.nf b/main.nf
@@ -23,7 +23,6 @@ include {
     mosdepth as mosdepth_input;
     mosdepth as mosdepth_downsampled;
     readStats;
-    mapula;
     getAllChromosomesBed;
     publish_artifact;
     configure_jbrowse;
@@ -217,14 +216,6 @@ workflow {
     } else {
         mosdepth_perbase = Channel.from("$projectDir/data/OPTIONAL_FILE")
     }
-
-    if (params.mapula) {
-        mapula(bam_channel, bed, ref_channel)
-        mapula_stats = mapula.out
-    }
-    else {
-        mapula_stats = Channel.empty()
-    }
 
     // Determine if the coverage threshold is met to perform analysis.
     // If too low, it creates an empty input channel, 
@@ -619,7 +610,6 @@ workflow {
             mosdepth_summary.flatten(),
             mosdepth_perbase.flatten(),
             mod_stats.flatten(),
-            mapula_stats.flatten(),
             jb_conf.flatten(),
             report_pass.flatten(),
             report_fail.flatten(),

diff --git a/modules/local/common.nf b/modules/local/common.nf
@@ -96,26 +96,6 @@ process mosdepth {
 }
 
 
-process mapula {
-    cpus 1
-    memory { 4.GB * task.attempt }
-    maxRetries 2
-    errorStrategy = {task.exitStatus in [137,140] ? 'retry' : 'finish'}
-    input:
-        tuple path(xam), path(xam_idx), val(xam_meta)
-        path target_bed
-        tuple path(ref), path(ref_idx), path(ref_cache), env(REF_PATH)
-    output:
-        tuple \
-            path("${params.sample_name}.mapula.csv"),
-            path("${params.sample_name}.mapula.json")
-    script:
-        """
-        mapula count -r ${ref} -f all -n '${params.sample_name}.mapula' ${xam}
-        """
-}
-
-
 process readStats {
     label "wf_common"
     cpus 4

diff --git a/nextflow.config b/nextflow.config
@@ -29,7 +29,6 @@ params {
     mod = false
     cnv = false
     str = false
-    mapula = false  // slow, should be deprecated
 
     // benchmark feature flags
     sv_benchmark = false
@@ -115,7 +114,7 @@ params {
         template_version = "195cab5"
         example_cmd = [
             "--bam 'wf-human-variation-demo/demo.bam'",
-            "--basecaller_cfg 'clair3:dna_r10.4.1_e8.2_400bps_hac_prom'",
+            "--basecaller_cfg 'dna_r10.4.1_e8.2_400bps_hac_prom'",
             "--mod",
             "--ref 'wf-human-variation-demo/demo.fasta'",
             "--sample_name 'DEMO'",
@@ -125,7 +124,7 @@ params {
         agent = null
 
         // container sha
-        e2l_base_tag = "sha0800eade05e4cbb75d45421633c78c4f6320b2f6"
+        e2l_base_tag = "sha26be394971f6444d657b0be6745a341beb0b2843"
         e2l_snp_tag = "sha14d7c4b75950fbffae7b2ea452dba21b263b84a1"
         e2l_sv_tag = "shabc3ac908a14705f248cdf49f218956ec33e93ef9"
         e2l_mod_tag = "shaeedb131a939d3eea2f9bd4dbecec805c0fa20bdb"
@@ -228,7 +227,7 @@ profiles {
             queue = { "${params.aws_queue}" }
             memory = '8G'
             withLabel:wfdefault {
-                container = { "${params.aws_image_prefix}-${params.wf.name}:${params.wf.e2l_base_tag}-root" }
+                container = { "${params.aws_image_prefix}-${params.wf.name}:${params.wf.e2l_base_tag}" }
             }
             withLabel:wf_human_snp {
                 container = { "${params.aws_image_prefix}-wf-human-variation-snp:${params.wf.e2l_snp_tag}-root" }

diff --git a/nextflow_schema.json b/nextflow_schema.json
@@ -49,13 +49,6 @@
                     "description": "Enable output of modified calls to a bedMethyl file [requires input BAM with Ml and Mm tags]",
                     "help_text": "This option is automatically selected and aggregation of modified calls with be carried out using modkit if Ml and Mm tags are found. Disable this option to prevent output of a bedMethyl file.",
                     "default": false
-                },
-                "mapula": {
-                    "type": "boolean",
-                    "description": "Enable mapula to generate basic alignment QC statistics (can be slow on very large data sets)",
-                    "help_text": "Mapula is a tool that is able to parse alignments and produce a range of useful stats. In most cases the other outputs are sufficient and this does not need to be enabled.",
-                    "hidden": true,
-                    "default": false
                 }
             },
             "anyOf": [