Initial commit: digital-patients pipeline (clean, no large files)

Large reference/model files excluded from repo - to be staged to S3 or baked into Docker images.

main_corto.nf

@@ -0,0 +1,46 @@
+nextflow.enable.dsl=2
+
+process CORTO {
+  memory 1.GB
+
+  container "${params.container_corto}"
+  containerOptions "${params.containerOptions}"
+  publishDir "${params.outdir}/${params.project_name}", mode: 'copy'
+  // debug true
+  // maxForks 1
+  stageInMode 'copy'
+
+  input:
+    path TPM
+    path regulon
+    // path TPM_ref
+
+  // Define output channels
+  output:
+    path "*_metabolome.csv", emit: csv_metabol
+
+  script:
+    """
+    #!/usr/bin/Rscript
+    library(corto)
+    library(data.table)
+
+    TPM <- as.matrix(fread("$TPM"),rownames=1)
+    #TPM_ref <- as.matrix(fread("TPM_ref"),rownames=1) #dollar sign is missing
+
+    #try normalization; dont use
+    #TPM <- scale(TPM)
+    #TPM_log2 = log2(TPM + 1)
+
+    load("$regulon")
+    
+    predicted<-mra(TPM, regulon=regulon)
+
+    #use this if we want to compare patients and reference TPM, data formet output is not csv -> work on data saving if needed
+    #predicted<-mra(TPM, path TPM_ref, regulon=regulon)
+
+    name = strsplit(strsplit("$TPM", split = "/")[[1]][length(strsplit("$TPM", split = "/")[[1]])],  split = "_TPM.csv")[[1]][1]
+    name = paste(name, "_metabolome.csv", sep="")
+    write.csv(predicted, name)
+    """
+}