Change GLIMPSE2_concordance non mandatory values (nf-core#7146)

* Update glimpse2 * Fix argument name * Fix nf-test * Improve glimpse2 concordance --------- Co-authored-by: Sateesh_Peri <[email protected]>
LouisLeNezet · Dec 4, 2024 · 6aed502 · 6aed502
1 parent 23600f6
commit 6aed502
Show file tree

Hide file tree

Showing 6 changed files with 65 additions and 180 deletions.
diff --git a/modules/nf-core/glimpse2/concordance/main.nf b/modules/nf-core/glimpse2/concordance/main.nf
@@ -9,9 +9,7 @@ process GLIMPSE2_CONCORDANCE {
 
     input:
     tuple val(meta), path(estimate), path(estimate_index), path(truth), path(truth_index), path(freq), path(freq_index), path(samples), val(region)
-    tuple val(meta2), path(groups), val(bins), val(ac_bins), val(allele_counts)
-    val(min_val_gl)
-    val(min_val_dp)
+    tuple val(meta2), path(groups), val(bins), val(ac_bins), val(allele_counts), val(min_val_gl), val(min_val_dp)
 
     output:
     tuple val(meta), path("*.error.cal.txt.gz")  , emit: errors_cal
@@ -26,17 +24,23 @@ process GLIMPSE2_CONCORDANCE {
     task.ext.when == null || task.ext.when
 
     script:
-    def args         = task.ext.args          ?: ''
-    def prefix       = task.ext.prefix        ?: "${meta.id}"
-    def samples_cmd  = samples                ? "--samples ${samples}"             : ""
-    def groups_cmd   = groups                 ? "--groups ${groups}"               : ""
-    def bins_cmd     = bins                   ? "--bins ${bins}"                   : ""
-    def ac_bins_cmd  = ac_bins                ? "--ac-bins ${ac_bins}"             : ""
-    def ale_ct_cmd   = allele_counts          ? "--allele-counts ${allele_counts}" : ""
-    def region_str   = region instanceof List ? region.join('\\n')                 : region
+    def args           = task.ext.args          ?: ''
+    def prefix         = task.ext.prefix        ?: "${meta.id}"
+    def samples_cmd    = samples                ? "--samples ${samples}"             : ""
+    def groups_cmd     = groups                 ? "--groups ${groups}"               : ""
+    def bins_cmd       = bins                   ? "--bins ${bins}"                   : ""
+    def ac_bins_cmd    = ac_bins                ? "--ac-bins ${ac_bins}"             : ""
+    def ale_ct_cmd     = allele_counts          ? "--allele-counts ${allele_counts}" : ""
+    def min_val_gl_cmd = min_val_gl             ? "--min-val-gl ${min_val_gl}"       : ""
+    def min_val_dp_cmd = min_val_dp             ? "--min-val-dp ${min_val_dp}"       : ""
+    def region_str     = region instanceof List ? region.join('\\n')                 : region
 
     if (((groups ? 1:0) + (bins ? 1:0) + (ac_bins ? 1:0) + (allele_counts ? 1:0)) != 1) error "One and only one argument should be selected between groups, bins, ac_bins, allele_counts"
 
+    if (args.contains("--gt-val")) {
+        assert !(min_val_gl || min_val_dp) : "If --gt-val is set, --min-val-gl nor --min-val-dp must be set"
+    }
+
     """
     printf '$region_str' > regions.txt
     sed 's/\$/ $freq $truth $estimate/' regions.txt > input.txt
@@ -47,8 +51,8 @@ process GLIMPSE2_CONCORDANCE {
         $bins_cmd \\
         $ac_bins_cmd \\
         $ale_ct_cmd \\
-        --min-val-gl $min_val_gl \\
-        --min-val-dp $min_val_dp \\
+        $min_val_gl_cmd \\
+        $min_val_dp_cmd \\
         --input input.txt \\
         --thread $task.cpus \\
         --output ${prefix}

diff --git a/modules/nf-core/glimpse2/concordance/meta.yml b/modules/nf-core/glimpse2/concordance/meta.yml
@@ -79,12 +79,12 @@ input:
         description: |
           Default allele count bins used for rsquared computations.
           AN field must be defined in the frequency file.
-  - - min_val_gl:
+    - min_val_gl:
         type: float
         description: |
           Minimum genotype likelihood probability P(G|R) in validation data.
           Set to zero to have no filter of if using –gt-validation
-  - - min_val_dp:
+    - min_val_dp:
         type: integer
         description: |
           Minimum coverage in validation data.

diff --git a/modules/nf-core/glimpse2/concordance/tests/main.nf.test b/modules/nf-core/glimpse2/concordance/tests/main.nf.test
@@ -10,78 +10,36 @@ nextflow_process {
     tag "modules_nfcore"
     tag "modules"
 
-    setup {
-        run("GLIMPSE2_PHASE") {
-            script "../../phase/main.nf"
-            process {
-                """
-                input_vcf     = Channel.of([
-                    [ id:'input'], // meta map
-                    file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/NA12878.chr21.s.1x.vcf.gz", checkIfExists: true),
-                    file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/NA12878.chr21.s.1x.vcf.gz.csi", checkIfExists: true),
-                    []
-                ])
-
-                samples_infos = Channel.of('NA12878 2').collectFile(name: 'sampleinfos.txt')
-                region        = Channel.of(["chr21:16600000-16800000","chr21:16650000-16750000"])
-
-                ch_ref_panel = Channel.of([
-                        file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/1000GP.chr21.noNA12878.s.bcf", checkIfExists: true),
-                        file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/1000GP.chr21.noNA12878.s.bcf.csi", checkIfExists: true)
-                ])
-
-                ch_map = Channel.of([
-                        file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/chr21.b38.gmap.gz", checkIfExists: true),
-                ])
-
-                // [meta, vcf, index, sample_infos, regionin, regionout, regionindex, ref, ref_index, map]
-                input[0] =  input_vcf
-                    .combine(samples_infos)
-                    .combine(region)
-                    .combine( ch_ref_panel )
-                    .combine( ch_map )
-                input[1]= Channel.of([[],[],[]])
-                """
-            }
-        }
-
-        run("BCFTOOLS_INDEX") {
-            script "../../../bcftools/index/main.nf"
-            process {
-                """
-                input[0] = GLIMPSE2_PHASE.output.phased_variants
-                """
-            }
-        }
-    }
-
-
-    test("test_glimpse2_concordance") {
+    test("test glimpse2 concordance") {
         config "./nextflow.config"
 
         when {
+            params {
+                glimpse2_concordance_args = "--gt-val --af-tag AF"
+            }
             process {
                 """
-                allele_freq = Channel.of([
-                    file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/1000GP.chr21.noNA12878.s.sites.vcf.gz",checkIfExists:true),
-                    file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/1000GP.chr21.noNA12878.s.sites.vcf.gz.csi",checkIfExists:true)
+                target = Channel.of([
+                    [id: "input"],
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878.chr22.1X.vcf.gz",checkIfExists:true),
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878.chr22.1X.vcf.gz.csi",checkIfExists:true)
                 ])
-
                 truth = Channel.of([
-                    file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/NA12878.chr21.s.bcf",checkIfExists:true),
-                    file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/NA12878.chr21.s.bcf.csi",checkIfExists:true)
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878_GIAB.chr22.vcf.gz",checkIfExists:true),
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878_GIAB.chr22.vcf.gz.csi",checkIfExists:true)
                 ])
-                list_inputs = GLIMPSE2_PHASE.output.phased_variants
-                        .join( BCFTOOLS_INDEX.out.csi )
+                allele_freq = Channel.of([
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/popgen/1000GP.chr22.sites.vcf.gz",checkIfExists:true),
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/popgen/1000GP.chr22.sites.vcf.gz.csi",checkIfExists:true)
+                ])
+                list_inputs = target
                         .combine( truth )
                         .combine( allele_freq )
                         .combine( Channel.of([[]]) )
-                        .combine( Channel.of(["chr21"]) )
+                        .combine( Channel.of(["chr22"]) )
 
                 input[0] = list_inputs
-                input[1] = Channel.of([[id:"params"],[],"0 0.01 0.05 0.1 0.2 0.5",[],[]])
-                input[2] = 0.9999
-                input[3] = 8
+                input[1] = Channel.of([[id:"params"], [],"0 0.01 0.05 0.1 0.2 0.5", [], [], [], []])
                 """
             }
         }
@@ -102,78 +60,37 @@ nextflow_process {
 
     }
 
-    test("test_list_region") {
+    test("test list of region and rsquare per site") {
         config "./nextflow.config"
 
         when {
+            params {
+                glimpse2_concordance_args = "--gt-val --af-tag AF --out-r2-per-site"
+            }
             process {
                 """
-                allele_freq = Channel.of([
-                    file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/1000GP.chr21.noNA12878.s.sites.vcf.gz",checkIfExists:true),
-                    file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/1000GP.chr21.noNA12878.s.sites.vcf.gz.csi",checkIfExists:true)
+                target = Channel.of([
+                    [id: "input"],
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878.chr22.1X.vcf.gz",checkIfExists:true),
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878.chr22.1X.vcf.gz.csi",checkIfExists:true)
                 ])
-
                 truth = Channel.of([
-                    file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/NA12878.chr21.s.bcf",checkIfExists:true),
-                    file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/NA12878.chr21.s.bcf.csi",checkIfExists:true)
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878_GIAB.chr22.vcf.gz",checkIfExists:true),
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/illumina/vcf/NA12878_GIAB.chr22.vcf.gz.csi",checkIfExists:true)
                 ])
-                list_inputs = GLIMPSE2_PHASE.output.phased_variants
-                        .join( BCFTOOLS_INDEX.out.csi )
-                        .combine( truth )
-                        .combine( allele_freq )
-                        .combine( Channel.of([[]]) )
-                        .combine( Channel.of(["chr21", "chr21"]) )
-
-                input[0] = list_inputs
-                input[1] = Channel.of([[id:"params"],[],"0 0.01 0.05 0.1 0.2 0.5",[],[]])
-                input[2] = 0.9999
-                input[3] = 8
-                """
-            }
-        }
-
-        then {
-            assertAll(
-                { assert process.success },
-                { assert snapshot(
-                    process.out.errors_cal.collect{ file(it[1]).name },
-                    process.out.errors_grp.collect{ file(it[1]).name },
-                    process.out.errors_spl.collect{ file(it[1]).name },
-                    process.out.rsquare_grp.collect{ file(it[1]).name },
-                    process.out.rsquare_spl.collect{ file(it[1]).name },
-                    process.out.versions
-                ).match()
-                }
-            )
-        }
-
-    }
-
-    test("test_r2_per_site") {
-        config "./nextflow_R2.config"
-
-        when {
-            process {
-                """
                 allele_freq = Channel.of([
-                    file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/1000GP.chr21.noNA12878.s.sites.vcf.gz",checkIfExists:true),
-                    file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/1000GP.chr21.noNA12878.s.sites.vcf.gz.csi",checkIfExists:true)
-                ])
-
-                truth = Channel.of([
-                    file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/NA12878.chr21.s.bcf",checkIfExists:true),
-                    file("https://github.com/nf-core/test-datasets/raw/modules/data/delete_me/glimpse/NA12878.chr21.s.bcf.csi",checkIfExists:true)
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/popgen/1000GP.chr22.sites.vcf.gz",checkIfExists:true),
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/popgen/1000GP.chr22.sites.vcf.gz.csi",checkIfExists:true)
                 ])
-                list_inputs = GLIMPSE2_PHASE.output.phased_variants
-                        .join( BCFTOOLS_INDEX.out.csi )
+                list_inputs = target
                         .combine( truth )
                         .combine( allele_freq )
                         .combine( Channel.of([[]]) )
-                        .combine( Channel.of(["chr21"]) )
+                        .combine( Channel.of([["chr22", "chr22"]]) )
+                        .view()
+
                 input[0] = list_inputs
-                input[1] = Channel.of([[id:"params"],[],"0 0.01 0.05 0.1 0.2 0.5",[],[]])
-                input[2] = 0.9999
-                input[3] = 8
+                input[1] = Channel.of([[id:"params"], [],"0 0.01 0.05 0.1 0.2 0.5", [], [], [], []])
                 """
             }
         }
@@ -189,11 +106,9 @@ nextflow_process {
                     process.out.rsquare_spl.collect{ file(it[1]).name },
                     process.out.rsquare_per_site.collect{ file(it[1]).name },
                     process.out.versions
-                ).match()
-                }
+                ).match() }
             )
         }
 
     }
-
 }
diff --git a/modules/nf-core/glimpse2/concordance/tests/main.nf.test.snap b/modules/nf-core/glimpse2/concordance/tests/main.nf.test.snap
@@ -1,5 +1,5 @@
 {
-    "test_r2_per_site": {
+    "test glimpse2 concordance": {
         "content": [
             [
                 "input.error.cal.txt.gz"
@@ -16,20 +16,17 @@
             [
                 "input.rsquare.spl.txt.gz"
             ],
-            [
-                "input_r2_sites.txt.gz"
-            ],
             [
                 "versions.yml:md5,ba729289bab6b9fbb8c36a620c86bb82"
             ]
         ],
         "meta": {
-            "nf-test": "0.8.4",
-            "nextflow": "24.04.3"
+            "nf-test": "0.9.1",
+            "nextflow": "24.10.1"
         },
-        "timestamp": "2024-10-22T16:23:00.623182365"
+        "timestamp": "2024-12-03T16:22:35.440086384"
     },
-    "test_glimpse2_concordance": {
+    "test list of region and rsquare per site": {
         "content": [
             [
                 "input.error.cal.txt.gz"
@@ -47,40 +44,16 @@
                 "input.rsquare.spl.txt.gz"
             ],
             [
-                "versions.yml:md5,ba729289bab6b9fbb8c36a620c86bb82"
-            ]
-        ],
-        "meta": {
-            "nf-test": "0.8.4",
-            "nextflow": "24.04.3"
-        },
-        "timestamp": "2024-10-22T16:22:26.823581573"
-    },
-    "test_list_region": {
-        "content": [
-            [
-                "input.error.cal.txt.gz"
-            ],
-            [
-                "input.error.grp.txt.gz"
-            ],
-            [
-                "input.error.spl.txt.gz"
-            ],
-            [
-                "input.rsquare.grp.txt.gz"
-            ],
-            [
-                "input.rsquare.spl.txt.gz"
+                "input_r2_sites.txt.gz"
             ],
             [
                 "versions.yml:md5,ba729289bab6b9fbb8c36a620c86bb82"
             ]
         ],
         "meta": {
-            "nf-test": "0.8.4",
-            "nextflow": "24.04.3"
+            "nf-test": "0.9.1",
+            "nextflow": "24.10.1"
         },
-        "timestamp": "2024-10-22T16:22:43.352729014"
+        "timestamp": "2024-12-03T16:26:35.009071185"
     }
 }
diff --git a/modules/nf-core/glimpse2/concordance/tests/nextflow.config b/modules/nf-core/glimpse2/concordance/tests/nextflow.config
@@ -1,6 +1,6 @@
 process {
-    withName: GLIMPSE2_CHUNK {
+    withName: GLIMPSE2_CONCORDANCE {
+        ext.args   = { params.glimpse2_concordance_args}
         ext.prefix = { "${meta.id}" }
     }
-    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
 }
diff --git a/modules/nf-core/glimpse2/concordance/tests/nextflow_R2.config b/modules/nf-core/glimpse2/concordance/tests/nextflow_R2.config