Updated EBSeq to v1.1.5

diff --git a/EBSeq/EBSeq_1.1.4.tar.gz b/EBSeq/EBSeq_1.1.4.tar.gz
deleted file mode 100644 (file)
index 4eb8059..0000000
Binary files a/EBSeq/EBSeq_1.1.4.tar.gz and /dev/null differ

diff --git a/EBSeq/makefile b/EBSeq/makefile
index ea5d2724b996a9ec2bdfeb1d6f742d705eaca797..020a32a292fb3ed2077cf388cba5bdcbf7ca92cf 100644 (file)
--- a/EBSeq/makefile
+++ b/EBSeq/makefile
@@ -6,8 +6,8 @@ all : $(PROGRAMS)
 blockmodeling : blockmodeling_0.1.8.tar.gz
        R CMD INSTALL -l "." blockmodeling_0.1.8.tar.gz
 
-EBSeq : blockmodeling EBSeq_1.1.4.tar.gz
-       R CMD INSTALL -l "." EBSeq_1.1.4.tar.gz
+EBSeq : blockmodeling EBSeq_1.1.5.tar.gz
+       R CMD INSTALL -l "." EBSeq_1.1.5.tar.gz
 
 rsem-for-ebseq-calculate-clustering-info : calcClusteringInfo.cpp
        $(CC) -O3 -Wall calcClusteringInfo.cpp -o $@

diff --git a/README.md b/README.md
index 4b0b3ec6bbea187f13a41dfeb8fb873aaa3b1d36..e59fb357ed34abab97a94ea5aaa1efe8e7bd25e8 100644 (file)
--- a/README.md
+++ b/README.md
@@ -379,18 +379,18 @@ Usage:
 
 This script calls EBSeq to find differentially expressed genes/transcripts in two conditions.
 
-data_matrix_file: m by n matrix containing expected counts, m is the number of transcripts/genes, n is the number of total samples.
-[--ngvector ngvector_file]: optional field. 'ngvector_file' is calculated by 'rsem-generate-ngvector'. Having this field is recommended for transcript data.
-number_sample_condition1: the number of samples in condition 1. A condition's samples must be adjacent. The left group of samples are defined as condition 1.
-FDR_rate: false discovery rate.
-output_file: the output file.
+data_matrix_file: m by n matrix containing expected counts, m is the number of transcripts/genes, n is the number of total samples.   
+[--ngvector ngvector_file]: optional field. 'ngvector_file' is calculated by 'rsem-generate-ngvector'. Having this field is recommended for transcript data.   
+number_sample_condition1: the number of samples in condition 1. A condition's samples must be adjacent. The left group of samples are defined as condition 1.   
+FDR_rate: false discovery rate.   
+output_file: the output file. Three files will be generated: 'output_file', 'output_file.hard_threshold' and 'output_file.all'. The first file reports all DE genes/transcripts using a soft threshold (calculated by crit_func in EBSeq). The second file reports all DE genes/transcripts using a hard threshold (only report if PPEE <= fdr). The third file reports all genes/transcripts. The first file is recommended to be used as DE results because it generally contains more called genes/transcripts.   
 
 The results are written as a matrix with row and column names. The row names are the differentially expressed transcripts'/genes' ids. The column names are 'PPEE', 'PPDE', 'PostFC' and 'RealFC'.
 
-PPEE: posterior probability of being equally expressed.
-PPDE: posterior probability of being differentially expressed.
-PostFC: posterior fold change (condition 1 over condition2).
-RealFC: real fold change (condition 1 over condition2).
+PPEE: posterior probability of being equally expressed.   
+PPDE: posterior probability of being differentially expressed.   
+PostFC: posterior fold change (condition 1 over condition2).   
+RealFC: real fold change (condition 1 over condition2).   
 
 To get the above usage information, type 
 

diff --git a/WHAT_IS_NEW b/WHAT_IS_NEW
index 31025d87755dbf889a055289557f69c1fae75de0..4e1690d3618661b9f3be45a9a36e2d21fad5ed38 100644 (file)
--- a/WHAT_IS_NEW
+++ b/WHAT_IS_NEW
@@ -1,3 +1,10 @@
+RSEM v1.2.2
+
+- Updated EBSeq to v1.1.5
+- Modified 'rsem-find-DE' to generate extra output files (type 'rsem-find-DE' to see more information)
+
+--------------------------------------------------------------------------------------------
+
 RSEM v1.2.1
 
 - Added poly(A) tails to 'reference_name.transcripts.fa' so that the RSEM generated transcript unsorted BAM file can be fed into RSEM as an input file. However, users need to rebuild their references if they want to visualize the transcript level wiggle files and BAM files using IGV

diff --git a/rsem-find-DE b/rsem-find-DE
index e9d25f7d38d957f2dbfc2e9a92762d16468d200f..e9d65cfbbd5fd0cb1a1446e0de1240b81e6339e1 100755 (executable)
--- a/rsem-find-DE
+++ b/rsem-find-DE
@@ -7,8 +7,8 @@ printUsage <- function() {
   cat("[--ngvector ngvector_file]: optional field. 'ngvector_file' is calculated by 'rsem-generate-ngvector'. Having this field is recommended for transcript data.\n")
   cat("number_sample_condition1: the number of samples in condition 1. A condition's samples must be adjacent. The left group of samples are defined as condition 1.\n")
   cat("FDR_rate: false discovery rate.\n")
-  cat("output_file: the output file.\n\n")
-  cat("The results are written as a matrix with row and column names. The row names are the differentially expressed transcripts'/genes' ids. The column names are 'PPEE', 'PPDE', 'PostFC' and 'RealFC'.\n\n")
+  cat("output_file: the output file. Three files will be generated: 'output_file', 'output_file.hard_threshold' and 'output_file.all'. The first file reports all DE genes/transcripts using a soft threshold (calculated by crit_func in EBSeq). The second file reports all DE genes/transcripts using a hard threshold (only report if PPEE <= fdr). The third file reports all genes/transcripts. The first file is recommended to be used as DE results because it generally contains more called genes/transcripts.\n\n")
+  cat("The results are written as a matrix with row and column names. The row names are the genes'/transcripts' ids. The column names are 'PPEE', 'PPDE', 'PostFC' and 'RealFC'.\n\n")
   cat("PPEE: posterior probability of being equally expressed.\n")
   cat("PPDE: posterior probability of being differentially expressed.\n")
   cat("PostFC: posterior fold change (condition 1 over condition2).\n")
@@ -59,11 +59,25 @@ if (is.null(ngvector)) {
 stopifnot(!is.null(EBOut))
 
 PP <- GetPPMat(EBOut)
+fc_res <- PostFC(EBOut)
+
+# soft threshold, default output
 thre <- crit_fun(PP[, "PPEE"], fdr)
 DEfound <- rownames(PP)[which(PP[, "PPDE"] >= thre)]
 
-fc_res <- PostFC(EBOut)
-
-results <- cbind(PP[DEfound, ], fc_res$GenePostFC[DEfound], fc_res$GeneRealFC[DEfound])
+results <- cbind(PP[DEfound, ], fc_res$PostFC[DEfound], fc_res$RealFC[DEfound])
 colnames(results) <- c("PPEE", "PPDE", "PostFC", "RealFC")
 write.table(results, file = output_file)
+
+# hard threshold
+thre <- 1.0 - fdr
+DEfound <- rownames(PP)[which(PP[, "PPDE"] >= thre)]
+
+results <- cbind(PP[DEfound, ], fc_res$PostFC[DEfound], fc_res$RealFC[DEfound])
+colnames(results) <- c("PPEE", "PPDE", "PostFC", "RealFC")
+write.table(results, file = paste(output_file, ".hard_threshold", sep = ""))
+
+# all
+results <- cbind(PP, fc_res$PostFC, fc_res$RealFC)
+colnames(results) <- c("PPEE", "PPDE", "PostFC", "RealFC")
+write.table(results, file = paste(output_file, ".all", sep = ""))