Add HDL h2 and rg modules

modules/nf-core/hdl/h2/environment.yml

@@ -0,0 +1,10 @@
+---
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/environment-schema.json
+channels:
+  - conda-forge
+  - bioconda
+  - dnachun
+dependencies:
+  - conda-forge::r-base=4.3.1
+  - conda-forge::r-data.table=1.15.4
+  - dnachun::r-hdl=1.4.0

modules/nf-core/hdl/h2/main.nf

@@ -0,0 +1,35 @@
+process HDL_H2 {
+    tag "hdl_h2_${meta.id}"
+    label 'process_medium'
+    conda "${moduleDir}/environment.yml"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'oras://community.wave.seqera.io/library/r-base_r-data.table_r-hdl:a2553504418194cc' :
+        'community.wave.seqera.io/library/r-base_r-data.table_r-hdl:cb9d70356e12d034' }"
+
+    input:
+    tuple val(meta), path(sumstats)
+    tuple val(meta2), path(hdl_ref_panel_dir)
+
+    output:
+    tuple val(meta), path("${meta.id}.h2.tsv"), emit: heritability_results
+    tuple val(meta), path("${meta.id}.hdl.log"), emit: h2_log
+    tuple val("${task.process}"), val("hdl"), val("1.4.0"), emit: versions_hdl, topic: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    template 'hdl_h2.R'
+
+    stub:
+    """
+    cat <<EOF_STUB > ${meta.id}.h2.tsv
+trait\th2\tse\tp\teigen_use
+${meta.id}\t0.2100\t0.0300\t0.00010\tautomatic
+EOF_STUB
+
+    cat <<EOF_STUB > ${meta.id}.hdl.log
+Stub HDL h2 run for ${meta.id}
+EOF_STUB
+    """
+}

modules/nf-core/hdl/h2/meta.yml

@@ -0,0 +1,62 @@
+name: hdl_h2
+description: Estimate SNP heritability from GWAS summary statistics using HDL.h2
+keywords:
+  - hdl
+  - heritability
+  - gwas
+tools:
+  - hdl:
+      description: High-definition likelihood method for genetic correlation and heritability from summary statistics
+      homepage: https://github.com/zhenin/HDL
+      documentation: https://github.com/zhenin/HDL/wiki
+input:
+  - - meta:
+        type: map
+        description: Metadata map for the trait input. The `meta.id` value is used to label output files.
+    - sumstats:
+        type: file
+        description: Canonicalized GWAS summary statistics for the trait in `meta.id`.
+  - - meta2:
+        type: map
+        description: Metadata map for the HDL reference panel directory.
+    - hdl_ref_panel_dir:
+        type: directory
+        description: Directory containing HDL LD reference panel files.
+output:
+  heritability_results:
+    - - meta:
+          type: map
+          description: Propagated trait metadata map.
+      - ${meta.id}.h2.tsv:
+          type: file
+          description: Tab-delimited HDL heritability summary output.
+  h2_log:
+    - - meta:
+          type: map
+          description: Propagated trait metadata map.
+      - ${meta.id}.hdl.log:
+          type: file
+          description: HDL heritability run log.
+  versions_hdl:
+    - - ${task.process}:
+          type: string
+          description: Name of the process
+      - hdl:
+          type: string
+          description: Name of the tool
+      - 1.4.0:
+          type: string
+          description: Version of the HDL R package bundled for the module
+topics:
+  versions:
+    - - ${task.process}:
+          type: string
+          description: Name of the process
+      - hdl:
+          type: string
+          description: Name of the tool
+      - 1.4.0:
+          type: string
+          description: Version of the HDL R package bundled for the module
+authors:
+  - andongni

modules/nf-core/hdl/h2/templates/hdl_h2.R

@@ -0,0 +1,41 @@
+#!/usr/bin/env Rscript
+
+suppressPackageStartupMessages({
+  library(data.table)
+})
+
+parse_hdl_ext_args <- function(raw_ext_args = "${task.ext.args ?: ''}") {
+  tryCatch(
+    eval(parse(text = sprintf("list(%s)", raw_ext_args))),
+    error = function(e) stop(sprintf("Failed to parse task.ext.args as R named arguments: %s", conditionMessage(e)))
+  )
+}
+
+if (!requireNamespace("HDL", quietly = TRUE)) {
+  stop("R package 'HDL' is not installed. Install HDL in the runtime container/environment.")
+}
+
+gwas <- fread("${sumstats}", data.table = FALSE)
+hdl_ext_args <- parse_hdl_ext_args()
+
+call_args <- c(
+  list(
+    gwas.df = gwas,
+    LD.path = "${hdl_ref_panel_dir}",
+    output.file = "${meta.id}.hdl.log"
+  ),
+  hdl_ext_args
+)
+
+result <- do.call(HDL::HDL.h2, call_args)
+
+out <- data.frame(
+  trait = "${meta.id}",
+  h2 = as.numeric(result[["h2"]]),
+  se = as.numeric(result[["h2.se"]]),
+  p = as.numeric(result[["P"]]),
+  eigen_use = as.character(result[["eigen.use"]]),
+  stringsAsFactors = FALSE
+)
+
+fwrite(out, "${meta.id}.h2.tsv", sep = "\t", quote = FALSE, na = "NA")

modules/nf-core/hdl/h2/tests/main.nf.test

@@ -0,0 +1,64 @@
+nextflow_process {
+
+    name "Test Process HDL_H2"
+    script "../main.nf"
+    config "./nextflow.config"
+    process "HDL_H2"
+    tag "modules"
+    tag "modules_nfcore"
+    tag "hdl"
+    tag "hdl/h2"
+
+    test("trait1 - real") {
+        when {
+            process {
+                """
+                input[0] = tuple(
+                    [id: "trait1"],
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/popgen/sumstats/trait1_canonical.tsv", checkIfExists: true)
+                )
+                input[1] = tuple(
+                    [id: "hdl_reference"],
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/popgen/hdl/reference", checkIfExists: true)
+                )
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert process.out.heritability_results.size() == 1 },
+                { assert process.out.h2_log.size() == 1 },
+                { assert process.out.versions_hdl.size() == 1 }
+            )
+        }
+    }
+
+    test("trait1 - stub") {
+        options "-stub"
+        when {
+            process {
+                """
+                input[0] = tuple(
+                    [id: "trait1"],
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/popgen/sumstats/trait1_canonical.tsv", checkIfExists: true)
+                )
+                input[1] = tuple(
+                    [id: "hdl_reference"],
+                    file(params.modules_testdata_base_path + "genomics/homo_sapiens/popgen/hdl/reference", checkIfExists: true)
+                )
+                """
+            }
+        }
+
+        then {
+            assertAll(
+                { assert process.success },
+                { assert process.out.heritability_results.size() == 1 },
+                { assert process.out.h2_log.size() == 1 },
+                { assert process.out.versions_hdl.size() == 1 }
+            )
+        }
+    }
+}

modules/nf-core/hdl/h2/tests/nextflow.config

@@ -0,0 +1,8 @@
+params {
+    modules_testdata_base_path = System.getenv("NF_MODULES_TESTDATA_BASE_PATH") ?: "https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/"
+}
+
+docker {
+    enabled    = true
+    runOptions = '-u $(id -u):$(id -g) --platform=linux/amd64'
+}

modules/nf-core/hdl/rg/environment.yml

@@ -0,0 +1,10 @@
+---
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/environment-schema.json
+channels:
+  - conda-forge
+  - bioconda
+  - dnachun
+dependencies:
+  - conda-forge::r-base=4.3.1
+  - conda-forge::r-data.table=1.15.4
+  - dnachun::r-hdl=1.4.0

modules/nf-core/hdl/rg/main.nf

@@ -0,0 +1,36 @@
+process HDL_RG {
+    tag "hdl_rg_${meta.id}_${meta2.id}"
+    label 'process_medium'
+    conda "${moduleDir}/environment.yml"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'oras://community.wave.seqera.io/library/r-base_r-data.table_r-hdl:a2553504418194cc' :
+        'community.wave.seqera.io/library/r-base_r-data.table_r-hdl:cb9d70356e12d034' }"
+
+    input:
+    tuple val(meta), path(sumstats1)
+    tuple val(meta2), path(sumstats2)
+    tuple val(meta3), path(hdl_ref_panel_dir)
+
+    output:
+    tuple val(meta), val(meta2), path("${meta.id}.${meta2.id}.rg.tsv"), emit: correlation_results
+    tuple val(meta), val(meta2), path("${meta.id}.${meta2.id}.hdl.log"), emit: rg_log
+    tuple val("${task.process}"), val("hdl"), val("1.4.0"), emit: versions_hdl, topic: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    template 'hdl_rg.R'
+
+    stub:
+    """
+    cat <<EOF_STUB > ${meta.id}.${meta2.id}.rg.tsv
+trait1\ttrait2\trg\tse\tp\th2_trait1\th2_trait2\tcovariance\teigen_use
+${meta.id}\t${meta2.id}\t0.3500\t0.0700\t0.00001\t0.2000\t0.1800\t0.0600\tautomatic
+EOF_STUB
+
+    cat <<EOF_STUB > ${meta.id}.${meta2.id}.hdl.log
+Stub HDL rg run for ${meta.id} and ${meta2.id}
+EOF_STUB
+    """
+}

modules/nf-core/hdl/rg/meta.yml

@@ -0,0 +1,74 @@
+name: hdl_rg
+description: Estimate pairwise genetic correlation from GWAS summary statistics using HDL.rg
+keywords:
+  - hdl
+  - genetic correlation
+  - gwas
+tools:
+  - hdl:
+      description: High-definition likelihood method for genetic correlation and heritability from summary statistics
+      homepage: https://github.com/zhenin/HDL
+      documentation: https://github.com/zhenin/HDL/wiki
+input:
+  - - meta:
+        type: map
+        description: Metadata map for the first trait input. The `meta.id` value is used as the first trait label.
+    - sumstats1:
+        type: file
+        description: Canonicalized GWAS summary statistics for the first trait.
+  - - meta2:
+        type: map
+        description: Metadata map for the second trait input. The `meta2.id` value is used as the second trait label.
+    - sumstats2:
+        type: file
+        description: Canonicalized GWAS summary statistics for the second trait.
+  - - meta3:
+        type: map
+        description: Metadata map for the HDL reference panel directory.
+    - hdl_ref_panel_dir:
+        type: directory
+        description: Directory containing HDL LD reference panel files.
+output:
+  correlation_results:
+    - - meta:
+          type: map
+          description: Propagated first-trait metadata map.
+      - meta2:
+          type: map
+          description: Propagated second-trait metadata map.
+      - ${meta.id}.${meta2.id}.rg.tsv:
+          type: file
+          description: Tab-delimited HDL genetic correlation summary output.
+  rg_log:
+    - - meta:
+          type: map
+          description: Propagated first-trait metadata map.
+      - meta2:
+          type: map
+          description: Propagated second-trait metadata map.
+      - ${meta.id}.${meta2.id}.hdl.log:
+          type: file
+          description: HDL genetic-correlation run log.
+  versions_hdl:
+    - - ${task.process}:
+          type: string
+          description: Name of the process
+      - hdl:
+          type: string
+          description: Name of the tool
+      - 1.4.0:
+          type: string
+          description: Version of the HDL R package bundled for the module
+topics:
+  versions:
+    - - ${task.process}:
+          type: string
+          description: Name of the process
+      - hdl:
+          type: string
+          description: Name of the tool
+      - 1.4.0:
+          type: string
+          description: Version of the HDL R package bundled for the module
+authors:
+  - andongni