From 8e136ce3cb46e31612efeee59342942912804dfa Mon Sep 17 00:00:00 2001
From: Jonathan Manning <jonathan.manning@seqera.io>
Date: Wed, 1 May 2024 14:06:51 +0100
Subject: [PATCH] Fix the PCA (#63)

* Fix the PCA

* fix up final assay
---
 R/pca.R                  | 4 ++--
 exec/exploratory_plots.R | 5 +++--
 2 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/R/pca.R b/R/pca.R
index 9538732..aff373f 100644
--- a/R/pca.R
+++ b/R/pca.R
@@ -279,9 +279,9 @@ runPCA <- function(matrix, do_log = TRUE) {
 compilePCAData <- function(matrix, ntop = NULL) {
 
   if (is.null(ntop)){
-    select <- selectVariableGenes(matrix = matrix, ntop = ntop)
-  }else{
     select <- 1:nrow(matrix)
+  }else{
+    select <- selectVariableGenes(matrix = matrix, ntop = ntop)
   }
 
   # perform a PCA on the data in assay(x) for the selected genes
diff --git a/exec/exploratory_plots.R b/exec/exploratory_plots.R
index 42e6332..efd8e6d 100755
--- a/exec/exploratory_plots.R
+++ b/exec/exploratory_plots.R
@@ -164,6 +164,7 @@ assay_files <-
     prettify_names = FALSE
   )
 
+
 # Valid samples are those with values in the specified sample metadata field
 
 valid_samples <- Filter(function(x) !is.na(x) && x != '' && !is.null(x), sample_metadata[[opt$contrast_variable]])
@@ -185,7 +186,7 @@ assay_data <- lapply(assay_files, function(x) {
 # Check an indicated final assay is among what we have
 
 if (is.null(opt$final_assay)){
-  final_assay <- length(assay_data)
+  final_assay <- names(assay_data)[length(assay_data)]
 }else{
   final_assay <- opt$final_assay
   if (!final_assay %in% names(assay_data)) {
@@ -272,7 +273,7 @@ dev.off()
 ################################################
 
 print("Writing PCA plots...")
-pca_data <- compilePCAData(assay_data[[final_assay]])
+pca_data <- compilePCAData(assay_data[[final_assay]], ntop = opt$n_genes)
 
 plotdata <- pca_data$coords
 plotdata$colorby <- factor(sample_metadata[[opt$contrast_variable]], levels = unique(sample_metadata[[opt$contrast_variable]]))