From 17654c9280b9be03eee626dfc1f813961e8ecb95 Mon Sep 17 00:00:00 2001
From: Keith Hughitt <keith.hughitt@gmail.com>
Date: Sat, 28 May 2016 08:17:07 -0400
Subject: [PATCH 1/2] Added check for rows in input dat with no variance

---
 R/ComBat.R | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/R/ComBat.R b/R/ComBat.R
index 116816b..b28fac5 100644
--- a/R/ComBat.R
+++ b/R/ComBat.R
@@ -38,6 +38,10 @@
 #' 
 
 ComBat <- function(dat, batch, mod=NULL, par.prior=TRUE,prior.plots=FALSE,mean.only=FALSE,ref.batch=NULL) {
+  # check to make sure there are no rows in the data with zero variance
+  if (sum(apply(dat, 1, var) == 0) > 0) {
+    stop("One or more rows in the data found to have zero variance.")
+  }
   # make batch a factor and make a set of indicators for batch
   if(mean.only==TRUE){cat("Using the 'mean only' version of ComBat\n")}
   if(length(dim(batch))>1){stop("This version of ComBat only allows one batch variable")}  ## to be updated soon!

From ca185e0600b49551f570956a92f2912f4d1b3253 Mon Sep 17 00:00:00 2001
From: Keith Hughitt <keith.hughitt@gmail.com>
Date: Sat, 28 May 2016 15:56:22 -0400
Subject: [PATCH 2/2] Updated ComBat function to ignore genes with zero
 variance when adjusting for batch

---
 R/ComBat.R | 22 +++++++++++++++++++---
 1 file changed, 19 insertions(+), 3 deletions(-)

diff --git a/R/ComBat.R b/R/ComBat.R
index b28fac5..aefd5ce 100644
--- a/R/ComBat.R
+++ b/R/ComBat.R
@@ -38,10 +38,19 @@
 #' 
 
 ComBat <- function(dat, batch, mod=NULL, par.prior=TRUE,prior.plots=FALSE,mean.only=FALSE,ref.batch=NULL) {
-  # check to make sure there are no rows in the data with zero variance
-  if (sum(apply(dat, 1, var) == 0) > 0) {
-    stop("One or more rows in the data found to have zero variance.")
+  # check for genes with uniform expression; if such genes exist, they will
+  # be temporarily removed from the data matrix before performing batch
+  # adjustment and add back in before returning the final result.
+  zero.var.rows <- apply(dat, 1, var) == 0
+
+  if (sum(zero.var.rows) > 0) {
+    cat(sprintf("Found %d genes with uniform expression; these will not be adjusted for batch.\n", sum(zero.var.rows)))
+
+    # keep a copy of the original data matrix and remove zero var rows
+    dat.orig <- dat
+    dat <- dat[!zero.var.rows,]
   }
+
   # make batch a factor and make a set of indicators for batch
   if(mean.only==TRUE){cat("Using the 'mean only' version of ComBat\n")}
   if(length(dim(batch))>1){stop("This version of ComBat only allows one batch variable")}  ## to be updated soon!
@@ -216,6 +225,13 @@ ComBat <- function(dat, batch, mod=NULL, par.prior=TRUE,prior.plots=FALSE,mean.o
   if(!is.null(ref.batch)){
     bayesdata[, batches[[ref]]] <- dat[, batches[[ref]]]
   }
+
+  # if genes genes with uniform expression were held out, add them back in now
+  if (sum(zero.var.rows) > 0) {
+    # keep a copy of the original data matrix and remove zero var rows
+    dat.orig[!zero.var.rows,] <- bayesdata
+    bayesdata <- dat.orig
+  }
   
   return(bayesdata)