@@ -1,5 +1,5 @@

 Package: GSVA

-Version: 1.25.1

+Version: 1.25.2

 Title: Gene Set Variation Analysis for microarray and RNA-seq data

 Authors@R: c(person("Justin", "Guinney", role=c("aut", "cre"), email="justin.guinney@sagebase.org"),

              person("Robert", "Castelo", role="aut", email="robert.castelo@upf.edu"))

@@ -463,11 +463,15 @@ compute.geneset.es <- function(expr, gset.idx.list, sample.idxs, rnaseq=FALSE,

 	}

 	rank.scores <- rep(0, num_genes)

-	if(abs.ranking){

-		sort.sgn.idxs <- apply(abs(gene.density), 2, order, decreasing=TRUE) # n.genes * n.samples	

-	}else{

-		sort.sgn.idxs <- apply(gene.density, 2, order, decreasing=TRUE) # n.genes * n.samples

-	}

+  ## 19.06.17 disable current functioning of abs.ranking flag

+  ## to switch to a different approach to implement it during

+  ## the KS random walk based on the Kuiper statistic

+	## if(abs.ranking){

+  ##   sort.sgn.idxs <- apply(abs(gene.density), 2, order, decreasing=TRUE) # n.genes * n.samples	

+  ## }else{

+  ##   sort.sgn.idxs <- apply(gene.density, 2, order, decreasing=TRUE) # n.genes * n.samples

+  ## }

+  sort.sgn.idxs <- apply(gene.density, 2, order, decreasing=TRUE) # n.genes * n.samples

 	rank.scores <- apply(sort.sgn.idxs, 2, compute_rank_score)

@@ -503,7 +507,8 @@ compute.geneset.es <- function(expr, gset.idx.list, sample.idxs, rnaseq=FALSE,

 		  m <- t(parSapp(cl, gset.idx.list, ks_test_m,

 						  gene.density=rank.scores, 

 						  sort.idxs=sort.sgn.idxs,

-						  mx.diff=mx.diff, tau=tau, verbose=FALSE))

+						  mx.diff=mx.diff, abs.ranking=abs.ranking,

+              tau=tau, verbose=FALSE))

 		  if(verbose)

         cat("Cleaning up\n")

 		  stopCl(cl)

@@ -528,7 +533,8 @@ compute.geneset.es <- function(expr, gset.idx.list, sample.idxs, rnaseq=FALSE,

       m <- mclapp(gset.idx.list, ks_test_m,

                   gene.density=rank.scores,

                   sort.idxs=sort.sgn.idxs,

-                  mx.diff=mx.diff, tau=tau, verbose=verbose)

+                  mx.diff=mx.diff, abs.ranking=abs.ranking,

+                  tau=tau, verbose=verbose)

       m <- do.call("rbind", m)

       colnames(m) <- colnames(expr)

@@ -555,7 +561,8 @@ compute.geneset.es <- function(expr, gset.idx.list, sample.idxs, rnaseq=FALSE,

     }

 		m <- t(sapply(gset.idx.list, ks_test_m, rank.scores, sort.sgn.idxs,

-                  mx.diff=mx.diff, tau=tau, verbose=verbose))

+                  mx.diff=mx.diff, abs.ranking=abs.ranking,

+                  tau=tau, verbose=verbose))

     if (verbose) {

       setTxtProgressBar(get("progressBar", envir=globalenv()), 1)

@@ -566,7 +573,8 @@ compute.geneset.es <- function(expr, gset.idx.list, sample.idxs, rnaseq=FALSE,

 }

-ks_test_m <- function(gset_idxs, gene.density, sort.idxs, mx.diff=TRUE, tau=1, verbose=TRUE){

+ks_test_m <- function(gset_idxs, gene.density, sort.idxs, mx.diff=TRUE,

+                      abs.ranking=FALSE, tau=1, verbose=TRUE){

 	n.genes <- nrow(gene.density)

 	n.samples <- ncol(gene.density)

@@ -581,7 +589,8 @@ ks_test_m <- function(gset_idxs, gene.density, sort.idxs, mx.diff=TRUE, tau=1, v

 			n.geneset,

 			as.double(tau),

 			n.samples,

-			as.integer(mx.diff))$R

+			as.integer(mx.diff),

+      as.integer(abs.ranking))$R

   if (verbose) {

     assign("iGeneSet", get("iGeneSet", envir=globalenv()) + 1, envir=globalenv())

@@ -103,10 +103,12 @@ Estimates GSVA enrichment scores.

                 RNA-seq counts such as log-CPMs, log-RPKMs or log-TPMs. This flag should be set to

                 \code{rnaseq=TRUE} only when the values of the input gene expression data are integer

                 counts.}

-  \item{abs.ranking}{Flag to determine whether genes should be ranked according to 

-  					their sign (\code{abs.ranking=FALSE}) or by absolute value (\code{abs.ranking=TRUE}). 

-  					In the latter, pathways with genes enriched on either extreme

-  					(high or low) will be regarded as 'highly' activated. }

+  \item{abs.ranking}{Flag used only when \code{mx.diff=TRUE}. When \code{abs.ranking=FALSE} (default)

+            a modified Kuiper statistic is used to calculate enrichment scores, taking the magnitude

+            difference between the largest positive and negative random walk deviations. When

+            \code{abs.ranking=TRUE} the original Kuiper statistic that sums the largest positive and

+            negative random walk deviations, is used. In this latter case, gene sets with genes

+            enriched on either extreme (high or low) will be regarded as 'highly' activated.}

   \item{min.sz}{Minimum size of the resulting gene sets.}

   \item{max.sz}{Maximum size of the resulting gene sets.}

   \item{no.bootstraps}{Number of bootstrap iterations to perform.}

@@ -203,7 +205,7 @@ geneSets <- list(set1=paste("g", 1:3, sep=""),

 y <- matrix(rnorm(n*p), nrow=p, ncol=n,

             dimnames=list(paste("g", 1:p, sep="") , paste("s", 1:n, sep="")))

-## genes in set1 are expressed at higher levels in the last 10 samples

+## genes in set1 are expressed at higher levels in the last 'nGrp1+1' to 'n' samples

 y[geneSets$set1, (nGrp1+1):n] <- y[geneSets$set1, (nGrp1+1):n] + 2

 ## build design matrix

@@ -6,10 +6,12 @@

 void

-ks_matrix_R(double* X, double* R, int* sidxs, int* n_genes, int* geneset_idxs, int* n_geneset, double* tau,  int* n_samples, int* mx_diff);

+ks_matrix_R(double* X, double* R, int* sidxs, int* n_genes, int* geneset_idxs,

+            int* n_geneset, double* tau,  int* n_samples, int* mx_diff, int* abs_rnk);

 double

-ks_sample(double* x, int* x_sort_indxs, int n_genes, int* geneset_mask, int* geneset_idxs, int n_geneset, double tau, int mx_diff){

+ks_sample(double* x, int* x_sort_indxs, int n_genes, int* geneset_mask,

+          int* geneset_idxs, int n_geneset, double tau, int mx_diff, int abs_rnk){

 	double dec = 1.0 / (n_genes - n_geneset);

@@ -40,9 +42,11 @@ ks_sample(double* x, int* x_sort_indxs, int n_genes, int* geneset_mask, int* gen

 		if(cum_sum < mx_neg){ mx_neg = cum_sum; }

 	}

-	if(mx_diff!=0){

+	if (mx_diff != 0) {

 		mx_value_sign = mx_pos + mx_neg;

-	}else{

+    if (abs_rnk != 0)

+      mx_value_sign = mx_pos - mx_neg;

+	} else {

 		mx_value_sign = (mx_pos > fabs(mx_neg)) ? mx_pos : mx_neg;

 	}

 	return mx_value_sign;

@@ -54,7 +58,8 @@ ks_sample(double* x, int* x_sort_indxs, int n_genes, int* geneset_mask, int* gen

  * R <- result

  * sidxs <- sorted gene densities idxs

  */

-void ks_matrix(double* X, double* R, int* sidxs, int n_genes, int* geneset_idxs, int n_geneset, double tau, int n_samples, int mx_diff){

+void ks_matrix(double* X, double* R, int* sidxs, int n_genes, int* geneset_idxs,

+               int n_geneset, double tau, int n_samples, int mx_diff, int abs_rnk){

 	int geneset_mask[n_genes];

 	for(int i = 0; i < n_genes; ++i){

 		geneset_mask[i] = 0;

@@ -66,11 +71,13 @@ void ks_matrix(double* X, double* R, int* sidxs, int n_genes, int* geneset_idxs,

 	for(int j = 0; j < n_samples; ++j){

 		int offset = j * n_genes;

-		R[j] = ks_sample(&X[offset], &sidxs[offset], n_genes, &geneset_mask[0], geneset_idxs, n_geneset, tau, mx_diff);

+    R[j] = ks_sample(&X[offset], &sidxs[offset], n_genes, &geneset_mask[0],

+                     geneset_idxs, n_geneset, tau, mx_diff, abs_rnk);

 	}

 }

-void ks_matrix_R(double* X, double* R, int* sidxs, int* n_genes, int* geneset_idxs, int* n_geneset, double* tau,  int* n_samples, int* mx_diff){

-	ks_matrix(X,R, sidxs, *n_genes, geneset_idxs, *n_geneset, *tau, *n_samples, *mx_diff);

+void ks_matrix_R(double* X, double* R, int* sidxs, int* n_genes,

+                 int* geneset_idxs, int* n_geneset, double* tau,

+                 int* n_samples, int* mx_diff, int* abs_rnk) {

+	ks_matrix(X,R, sidxs, *n_genes, geneset_idxs, *n_geneset, *tau, *n_samples, *mx_diff, *abs_rnk);

 }

-

@@ -9,7 +9,7 @@ void

 matrix_density_R(double* X, double* Y, double* R, int* n_density_samples,int* n_test_samples, int* n_genes, int* rnaseq);

 void

-ks_matrix_R(double* X, double* R, int* sidxs, int* n_genes, int* geneset_idxs, int* n_geneset, double* tau,  int* n_samples, int* mx_diff);

+ks_matrix_R(double* X, double* R, int* sidxs, int* n_genes, int* geneset_idxs, int* n_geneset, double* tau,  int* n_samples, int* mx_diff, int* abs_rnk);

 /* registration of C-entry points */

@@ -17,12 +17,12 @@ static R_NativePrimitiveArgType

 matrix_density_R_t[7] = {REALSXP, REALSXP, REALSXP, INTSXP, INTSXP, INTSXP, INTSXP};

 static R_NativePrimitiveArgType

-ks_matrix_R_t[9] = {REALSXP, REALSXP, INTSXP, INTSXP, INTSXP, INTSXP, REALSXP, INTSXP, INTSXP};

+ks_matrix_R_t[10] = {REALSXP, REALSXP, INTSXP, INTSXP, INTSXP, INTSXP, REALSXP, INTSXP, INTSXP, INTSXP};

 static const R_CMethodDef

 cMethods[] = {

   {"matrix_density_R", (DL_FUNC) &matrix_density_R, 7, matrix_density_R_t},

-  {"ks_matrix_R", (DL_FUNC) &ks_matrix_R, 9, ks_matrix_R_t},

+  {"ks_matrix_R", (DL_FUNC) &ks_matrix_R, 10, ks_matrix_R_t},

   {NULL, NULL, 0}

 };