fix switching between fraser versions to start from right script

gagneurlab · Apr 13, 2023 · 54d157b · 54d157b
1 parent 29663fa
commit 54d157b
Show file tree

Hide file tree

Showing 7 changed files with 53 additions and 14 deletions.
diff --git a/drop/modules/aberrant-splicing-pipeline/Counting/02_psi_value_calculation_FraseR.R b/drop/modules/aberrant-splicing-pipeline/Counting/02_psi_value_calculation_FraseR.R
@@ -12,8 +12,8 @@
 #'   - counting_done: '`sm cfg.getProcessedDataDir() + 
 #'                "/aberrant_splicing/datasets/savedObjects/raw-local-{dataset}/counting.done" `'
 #'  output:
-#'  - theta:     '`sm cfg.getProcessedDataDir() +
-#'                    "/aberrant_splicing/datasets/savedObjects/raw-local-{dataset}/theta.h5"`'
+#'  - splice_metrics: '`sm expand(cfg.getProcessedDataDir() +
+#'                    "/aberrant_splicing/datasets/savedObjects/raw-local-{dataset}/{type}.h5", type=cfg.AS.getPsiTypeAssay(), allow_missing=True)`'
 #'  type: script
 #'--- 
 

diff --git a/drop/modules/aberrant-splicing-pipeline/Counting/03_filter_expression_FraseR.R b/drop/modules/aberrant-splicing-pipeline/Counting/03_filter_expression_FraseR.R
@@ -9,14 +9,14 @@
 #'   - workingDir: '`sm cfg.getProcessedDataDir() + "/aberrant_splicing/datasets/"`'
 #'   - exCountIDs: '`sm lambda w: sa.getIDsByGroup(w.dataset, assay="SPLICE_COUNT")`'
 #'  input:
-#'   - theta:  '`sm cfg.getProcessedDataDir() +
-#'                  "/aberrant_splicing/datasets/savedObjects/raw-local-{dataset}/theta.h5"`'
+#'   - splice_metrics: '`sm expand(cfg.getProcessedDataDir() +
+#'                  "/aberrant_splicing/datasets/savedObjects/raw-local-{dataset}/{type}.h5", type=cfg.AS.getPsiTypeAssay(), allow_missing=True)`'
 #'   - exCounts: '`sm lambda w: cfg.AS.getExternalCounts(w.dataset, "k_j_counts")`'
 #'  output:
 #'   - fds: '`sm cfg.getProcessedDataDir() +
 #'                  "/aberrant_splicing/datasets/savedObjects/{dataset}/fds-object.RDS"`'
-#'   - done: '`sm cfg.getProcessedDataDir() +
-#'                  "/aberrant_splicing/datasets/savedObjects/{dataset}/filter.done" `'
+#'   - done: '`sm expand(cfg.getProcessedDataDir() +
+#'                  "/aberrant_splicing/datasets/savedObjects/{dataset}/filter_{version}.done", version=cfg.AS.get("FRASER_version"), allow_missing=True)`'
 #'  threads: 3
 #'  type: script
 #'---
@@ -97,4 +97,9 @@ if (params$filter == TRUE) {
 seqlevels(fds) <- seqlevelsInUse(fds)
 colData(fds)$sampleID <- as.character(colData(fds)$sampleID)
 fds <- saveFraserDataSet(fds,dir = workingDir)
+
+# remove previous filter.done files and create new one
+outdir <- dirname(snakemake@output$done)
+prevFilterFiles <- grep("filter(.*)done", list.files(outdir), value=TRUE)
+unlink(file.path(outdir, prevFilterFiles))
 file.create(snakemake@output$done)
diff --git a/drop/modules/aberrant-splicing-pipeline/FRASER/04_fit_hyperparameters_FraseR.R b/drop/modules/aberrant-splicing-pipeline/FRASER/04_fit_hyperparameters_FraseR.R
@@ -9,11 +9,11 @@
 #'   - workingDir: '`sm cfg.getProcessedDataDir() + "/aberrant_splicing/datasets/"`'
 #'  threads: 12
 #'  input:
-#'   - filter: '`sm cfg.getProcessedDataDir() + 
-#'                "/aberrant_splicing/datasets/savedObjects/{dataset}/filter.done" `'
+#'   - filter: '`sm expand(cfg.getProcessedDataDir() + 
+#'                "/aberrant_splicing/datasets/savedObjects/{dataset}/filter_{version}.done", version=cfg.AS.get("FRASER_version"), allow_missing=True)`'
 #'  output:
-#'   - hyper: '`sm cfg.getProcessedDataDir() + 
-#'                "/aberrant_splicing/datasets/savedObjects/{dataset}/hyper.done" `'
+#'   - hyper: '`sm expand(cfg.getProcessedDataDir() + 
+#'                "/aberrant_splicing/datasets/savedObjects/{dataset}/hyper_{version}.done", version=cfg.AS.get("FRASER_version"), allow_missing=True)`'
 #'  type: script
 #'---
 
@@ -66,4 +66,9 @@ for(type in psiTypes){
     fds <- saveFraserDataSet(fds)
 }
 fds <- saveFraserDataSet(fds)
+
+# remove previous hyper.done files and create new one
+outdir <- dirname(snakemake@output$hyper)
+prevFilterFiles <- grep("hyper(.*)done", list.files(outdir), value=TRUE)
+unlink(file.path(outdir, prevFilterFiles))
 file.create(snakemake@output$hyper)
diff --git a/drop/modules/aberrant-splicing-pipeline/FRASER/05_fit_autoencoder_FraseR.R b/drop/modules/aberrant-splicing-pipeline/FRASER/05_fit_autoencoder_FraseR.R
@@ -9,8 +9,8 @@
 #'   - workingDir: '`sm cfg.getProcessedDataDir() + "/aberrant_splicing/datasets/"`'
 #'  threads: 20
 #'  input:
-#'   - hyper: '`sm cfg.getProcessedDataDir() + 
-#'                "/aberrant_splicing/datasets/savedObjects/{dataset}/hyper.done" `'
+#'   - hyper: '`sm expand(cfg.getProcessedDataDir() + 
+#'                "/aberrant_splicing/datasets/savedObjects/{dataset}/hyper_{version}.done", version=cfg.AS.get("FRASER_version"), allow_missing=True)`'
 #'  output:
 #'   - fdsout: '`sm expand(cfg.getProcessedDataDir() + 
 #'                  "/aberrant_splicing/datasets/savedObjects/{dataset}/predictedMeans_{type}.h5", type=cfg.AS.getPsiTypeAssay(), allow_missing=True)`'
@@ -41,3 +41,11 @@ for(type in psiTypes){
     fds <- saveFraserDataSet(fds)
 }
 
+# remove .h5 files from previous runs with other FRASER version
+fdsDir <- dirname(snakemake@output$fdsout[1])
+for(type in psiTypesNotUsed){
+    predMeansFile <- file.path(fdsDir, paste0("predictedMeans_", type, ".h5"))
+    if(file.exists(predMeansFile)){
+        unlink(predMeansFile)
+    }
+}
diff --git a/drop/modules/aberrant-splicing-pipeline/FRASER/06_annotate_genes.R b/drop/modules/aberrant-splicing-pipeline/FRASER/06_annotate_genes.R
@@ -60,3 +60,11 @@ fds_input <- annotateIntronReferenceOverlap(fds_input, txdb)
 # save fds
 fds <- saveFraserDataSet(fds_input, dir=outputDir, name = paste(dataset, annotation, sep = '--'), rewrite = TRUE)
 
+# remove .h5 files from previous runs with other FRASER version
+fdsDir <- dirname(snakemake@output$fdsout[1])
+for(type in psiTypesNotUsed){
+    predMeansFile <- file.path(fdsDir, paste0("predictedMeans_", type, ".h5"))
+    if(file.exists(predMeansFile)){
+        unlink(predMeansFile)
+    }
+}
diff --git a/drop/modules/aberrant-splicing-pipeline/FRASER/07_calculation_stats_FraseR.R b/drop/modules/aberrant-splicing-pipeline/FRASER/07_calculation_stats_FraseR.R
@@ -53,3 +53,14 @@ for (type in psiTypes) {
 
 fds <- saveFraserDataSet(fds)
 
+# remove .h5 files from previous runs with other FRASER version
+fdsDir <- dirname(snakemake@output$fdsout[1])
+pvalFiles <- grep("p(.*)BetaBinomial_(.*).h5", 
+                  list.files(fdsDir), 
+                  value=TRUE)
+for(type in psiTypesNotUsed){
+    pvalFilesType <- grep(type, pvalFiles, value=TRUE)
+    for(pFile in pvalFilesType){
+        unlink(file.path(fdsDir, pFile))
+    }
+}
diff --git a/drop/modules/aberrant-splicing-pipeline/config.R b/drop/modules/aberrant-splicing-pipeline/config.R
@@ -35,9 +35,11 @@ h5disableFileLocking()
 # set psiTypes to run based on preference in config.yaml
 cfg <- yaml::read_yaml("config.yaml")
 if(cfg$aberrantSplicing$FRASER_version == "FRASER2"){
-    psiTypes <- c("jaccard")
     pseudocount(0.1)
+    psiTypes <- c("jaccard")
+    psiTypesNotUsed <- c("psi5", "psi3", "theta")
 } else{
-    psiTypes <- c("psi5", "psi3", "theta")
     pseudocount(1)
+    psiTypes <- c("psi5", "psi3", "theta")
+    psiTypesNotUsed <- c("jaccard")
 }