Update to package

DESCRIPTION

@@ -1,5 +1,5 @@
 Package: select
-Version: 1.6.3
+Version: 1.6.4
 Date: 2019-12-01
 Title: SELECT - Selected Events Linked by Evolutionary Conditions across human Tumors
 Authors@R: c(  

NEWS.md

@@ -1,3 +1,7 @@
+# select 1.6.4
+
+* Fixed some bugs and updated code.
+
 # select 1.6.3
 
 * Added a `NEWS.md` file to track changes to the package.

R/select.R

@@ -42,7 +42,7 @@
 	calculate_FDR = TRUE,
 	verbose = TRUE,
 	nbatches = 1,
-	FDR.thresh = 0.1
+	FDR.thresh = FDR.cutoff
 	) {
 	print(paste('-> Collecting event stats on observed GAM...'))
 	# Get AL stats
@@ -166,20 +166,25 @@
 			print(paste('Memory usage:', gc()['Vcells',2], 'MB'))
 		}
 	}
-
+	if(save.intermediate.files) save(FDR, file=paste(folder,'wMI_FDR.RData',sep=''))
+
 	alpi$select_score_good_cancer_cell_2017_criterion_1 = NA
 	# Select optimal APC score threshold (removing 95% of FP)
 	if(calculate_APC_threshold) {
 		print(paste('-> Calculating APC threshold...'))
 		TP.FP.thr = select:::establish_APC_threshold(alpi)
 		threshold = TP.FP.thr$thresh['t']
+		#print(threshold)
 		alpi$select_score_good_cancer_cell_2017_criterion_1 = FALSE
 		alpi$select_score_good_cancer_cell_2017_criterion_1[which(alpi$APC >= threshold)] = TRUE
+		#print(dim(alpi))
+		#print(colnames(alpi))
 	}
 
+	#print(head(alpi,3))
 	alpi$select_score = alpi$APC
-	alpi$APC = NULL
-
+	#alpi$APC = NULL
+	
 	# Save pairwise results
 	if(save.intermediate.files) save(alpi, file=paste(folder,'alpi.RData',sep=''))
 	out.format = c('name','SFE_1','SFE_2','int_type', 'support_1', 'support_2', 'overlap', 'r_overlap', 'max_overlap', 'direction', 'freq_overlap', 'wMI_p.value', 'ME_p.value', 'wMI_p.value_FDR', 'select_score', 'select_score_good_cancer_cell_2017_criterion_1')
@@ -238,6 +243,7 @@ select <- function(
 	save.intermediate.files = TRUE,
 	randomization.switch.threshold = 30,
 	max.memory.size=100,
+	FDR.cutoff=0.1,
 	calculate_APC_threshold = TRUE,
 	calculate_FDR = TRUE,
 	verbose = TRUE
@@ -296,6 +302,7 @@ select <- function(
 		save.intermediate.files = save.intermediate.files,
 		calculate_APC_threshold = calculate_APC_threshold,
 		calculate_FDR = calculate_FDR,
+		FDR.thresh=FDR.cutoff,
 		verbose = verbose,
 		nbatches = nbatches
 	)

R/select_APC_threshold.R

@@ -27,26 +27,29 @@ TP.FP.stats <- function(A,res=1000) {
 	AA$FP = 1 - AA$TN
 	AA$F1 = 2 * AA$TP / (2 * AA$TP + AA$FN + AA$FP)
 	AA$MCC = ((AA$TP * AA$TN) - (AA$FP * AA$FN)) / sqrt((AA$TP + AA$FP)*(AA$TP + AA$FN)*(AA$TN + AA$FP)*(AA$TN + AA$FN))
-	return(as.vector(AA))
+	return(AA)
 }
 
 #' Establish which threshold is the best, using the 95\% of FP exclusion criterion
 get_thresh.2 <- function(x, A) {
+	#print('here')
+	#print(dim(x))
+	#print(x[,'FP'])
 	TP.FP.thresh_v1 = x[x[,'FP'] <= 0.05,,drop=FALSE]
 	TP.FP.thresh_v2 = x[x[,'FP'] >= 0.05,,drop=FALSE]
 	if(is.null(x)) return(NA) # there are no solutions at all. return NA
 	if(is.null(dim(x))) return(NA) # there are no solutions at all. return NA
 	if(nrow(x)==0) return(NA) # there are no solutions at all. return NA
 	if(nrow(TP.FP.thresh_v1)==0) return(x[nrow(x),]) # there are no good solutions. Set highest th possible 
 	if(nrow(TP.FP.thresh_v2)==0) return(TP.FP.thresh_v1[1,]) # All solutions are good. Set threshold to minimum score of good solutions
-	# print('here')
-	# print(TP.FP.thresh_v1[1,'t'])
+	#print('here')
+	#print(TP.FP.thresh_v1[1,'t'])
 	# print(TP.FP.thresh_v2[nrow(TP.FP.thresh_v2),'t'])
 	th = (TP.FP.thresh_v1[1,'t'] + TP.FP.thresh_v2[nrow(TP.FP.thresh_v2),'t']) / 2
 	th = (TP.FP.thresh_v1[1,'t']) # + TP.FP.thresh_v2[nrow(TP.FP.thresh_v2),'t']) / 2
 	th = TP.FP.thresh_v2[nrow(TP.FP.thresh_v2),'t'] # Set threshold at maximum score associated to FP > 0.05
-	# print(th)
-	# print(TP.FP.stat(A, th))
+	#print(th)
+	#print(TP.FP.stat(A, th))
 	return(TP.FP.stat(A, th))
 	# return(TP.FP.thresh_v1[1,])
 }
@@ -63,7 +66,10 @@ establish_APC_threshold <- function(alpi) {
 		thresh = rep(NA, 7)
 	} else {
 		stats = TP.FP.stats(alpi, 10000)
+		#print(str(stats))
+		#print(head(alpi,2))
 		thresh = get_thresh.2(stats, alpi)
+		#print(str(thresh))
 	}
 	return(list('stats' = stats, thresh = thresh))
 

README.md

@@ -5,7 +5,7 @@ This is the official repository of the R package SELECT - Selected Events Linked
 ### What is this repository for? ###
 
 * SELECT is a package for inferring evolutionary dependencies between functional alterations in cancer starting from patterns of alteration occurrences in large tumor sample cohorts.
-* Last Version: 1.6.3
+* Last Version: 1.6.4
 
 
 ### How To ###
@@ -19,18 +19,19 @@ devtools::install_github("CSOgroup/select")
 
 * How to use
 	- Check the introduction fom vignettes folder or the website.
+	- Also refere to *methods section* of papers to know the best way to apply the method.
 
 ### Citation ###
 
 If you use this tool, please cite the papers
 
-- Mina, M., Iyer, A., Tavernari, D., Raynaud, F., & Ciriello, G. (2020). Discovering functional evolutionary dependencies in human cancers. Nature Genetics, 52(11), 1198-1207.
-- Mina, M., Raynaud, F., Tavernari, D., Battistello, E., Sungalee, S., Saghafinia, S., ... & Ciriello, G. (2017). Conditional selection of genomic alterations dictates cancer evolution and oncogenic dependencies. Cancer cell, 32(2), 155-168.
+- Mina, M., Iyer, A., Tavernari, D., Raynaud, F., & Ciriello, G. (2020). [Discovering functional evolutionary dependencies in human cancers](https://www.nature.com/articles/s41588-020-0703-5) Nature Genetics, 52(11), 1198-1207. 
+- Mina, M., Raynaud, F., Tavernari, D., Battistello, E., Sungalee, S., Saghafinia, S., ... & Ciriello, G. (2017). [Conditional selection of genomic alterations dictates cancer evolution and oncogenic dependencies.](https://www.cell.com/cancer-cell/fulltext/S1535-6108(17)30261-1) Cancer cell, 32(2), 155-168.
 
 ### Contribution guidelines ###
 
-- SELECT was originally developed by Marco Mina at the University of Lausanne, Switzerland.
-- SELECT is currently maintained by Arvind Iyer from the Computational Systems Oncology lab (CSO) at the University of Lausanne (Unil), Switzerland.
+- SELECT was originally developed by **Marco Mina** at the University of Lausanne, Switzerland.
+- SELECT is currently maintained by Arvind Iyer and memeber of the Computational Systems Oncology lab (CSO) at the University of Lausanne (Unil), Switzerland.
 
 ### Who do I talk to? ###