From 7efa8d83bcb0febf9265b6e2d4d470477273a794 Mon Sep 17 00:00:00 2001
From: TuomasBorman <tvborm@utu.fi>
Date: Fri, 22 Nov 2024 20:24:13 +0200
Subject: [PATCH 1/3] Add main nf file for TreeSE

---
 modules/local/treesummarizedexperiment.nf | 72 +++++++++++++++++++++++
 1 file changed, 72 insertions(+)
 create mode 100644 modules/local/treesummarizedexperiment.nf

diff --git a/modules/local/treesummarizedexperiment.nf b/modules/local/treesummarizedexperiment.nf
new file mode 100644
index 00000000..512426db
--- /dev/null
+++ b/modules/local/treesummarizedexperiment.nf
@@ -0,0 +1,72 @@
+process TREESUMMARIZEDEXPERIMENT {
+    tag "$prefix"
+    label 'process_low'
+
+    conda "bioconda::bioconductor-treesummarizedexperiment=2.10.0"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/bioconductor-treesummarizedexperiment%3A2.8.0--r43hdfd78af_0' :
+        'bioconductor-treesummarizedexperiment%3A2.8.0--r43hdfd78af_0' }"
+
+    input:
+    tuple val(prefix), path(tax_tsv), path(otu_tsv)
+    path sam_tsv
+    path tree
+
+    output:
+    tuple val(prefix), path("*TreeSummarizedExperiment.rds"), emit: rds
+    path "versions.yml"                     , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def sam_tsv = "\"${sam_tsv}\""
+    def otu_tsv = "\"${otu_tsv}\""
+    def tax_tsv = "\"${tax_tsv}\""
+    def tree    = "\"${tree}\""
+    def prefix  = "\"${prefix}\""
+    """
+    #!/usr/bin/env Rscript
+
+    suppressPackageStartupMessages(library(TreeSummarizedExperiment))
+
+    # Read otu table. It must be in a list as a matrix.
+    otu_mat  <- read.table($otu_tsv, sep="\\t", header=TRUE, row.names=1)
+    otu_mat <- as.matrix(otu_mat)
+    assays <- SimpleList(counts = otu_mat)
+    # Read taxonomy table. Correct format for it is DataFrame.
+    taxonomy_table  <- read.table($tax_tsv, sep="\\t", header=TRUE, row.names=1)
+    taxonomy_table <- DataFrame(taxonomy_table)
+
+    # Create TreeSE object. We assume that rownames between taxonomy table
+    # and abundance matrix match.
+    tse <- TreeSummarizedExperiment(
+        assays = assays,
+        rowData = taxonomy_table
+    )
+
+    # If provided, we add sample metadata as DataFrame object. rownames of
+    # sample metadata must match with colnames of abundance matrix.
+    if (file.exists($sam_tsv)) {
+        sample_meta  <- read.table($sam_tsv, sep="\\t", header=TRUE, row.names=1)
+        sample_meta  <- DataFrame(sample_meta)
+        colData(tse) <- sample_meta
+    }
+
+    # If provided, we add phylogeny. The rownames in abundance matrix must match
+    # with node labels in phylogeny.
+    if (file.exists($tree)) {
+        phylogeny <- read_tree($tree)
+        rowTree(tse) <- phylogeny
+    }
+
+    saveRDS(tse, file = paste0($prefix, "_TreeSummarizedExperiment.rds"))
+
+    # Version information
+    writeLines(c("\\"${task.process}\\":",
+        paste0("    R: ", paste0(R.Version()[c("major","minor")], collapse = ".")),
+        paste0("    TreeSummarizedExperiment: ", packageVersion("TreeSummarizedExperiment"))),
+        "versions.yml"
+    )
+    """
+}

From fbd3bfab876d4f62c801561fce4a203d4712a206 Mon Sep 17 00:00:00 2001
From: TuomasBorman <tvborm@utu.fi>
Date: Fri, 22 Nov 2024 20:31:47 +0200
Subject: [PATCH 2/3] up

---
 modules/local/treesummarizedexperiment.nf | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/modules/local/treesummarizedexperiment.nf b/modules/local/treesummarizedexperiment.nf
index 512426db..7a3ec269 100644
--- a/modules/local/treesummarizedexperiment.nf
+++ b/modules/local/treesummarizedexperiment.nf
@@ -30,7 +30,8 @@ process TREESUMMARIZEDEXPERIMENT {
 
     suppressPackageStartupMessages(library(TreeSummarizedExperiment))
 
-    # Read otu table. It must be in a list as a matrix.
+    # Read otu table. It must be in a SimpleList as a matrix where rows
+    # represent taxa and columns samples.
     otu_mat  <- read.table($otu_tsv, sep="\\t", header=TRUE, row.names=1)
     otu_mat <- as.matrix(otu_mat)
     assays <- SimpleList(counts = otu_mat)

From bddb10434297f8ef3cd85bdd73321c404a9d8ae4 Mon Sep 17 00:00:00 2001
From: TuomasBorman <tvborm@utu.fi>
Date: Fri, 22 Nov 2024 20:46:47 +0200
Subject: [PATCH 3/3] Fix TreeSE version

---
 modules/local/treesummarizedexperiment.nf | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/modules/local/treesummarizedexperiment.nf b/modules/local/treesummarizedexperiment.nf
index 7a3ec269..d72b05fa 100644
--- a/modules/local/treesummarizedexperiment.nf
+++ b/modules/local/treesummarizedexperiment.nf
@@ -4,8 +4,8 @@ process TREESUMMARIZEDEXPERIMENT {
 
     conda "bioconda::bioconductor-treesummarizedexperiment=2.10.0"
     container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
-        'https://depot.galaxyproject.org/singularity/bioconductor-treesummarizedexperiment%3A2.8.0--r43hdfd78af_0' :
-        'bioconductor-treesummarizedexperiment%3A2.8.0--r43hdfd78af_0' }"
+        'https://depot.galaxyproject.org/singularity/bioconductor-treesummarizedexperiment%3A2.10.0--r43hdfd78af_0' :
+        'bioconductor-treesummarizedexperiment%3A2.10.0--r43hdfd78af_0' }"
 
     input:
     tuple val(prefix), path(tax_tsv), path(otu_tsv)