Changed the behavior of parsing GTF file. Now if a GTF line's feature is not exon...

diff --git a/GTFItem.h b/GTFItem.h
index a908375c9296ca1ba9ce0717969b11e30bffd109..e0499be9139ab226a1643aea47e379e92af275f4 100644 (file)
--- a/GTFItem.h
+++ b/GTFItem.h
@@ -80,8 +80,10 @@ public:
                        }
                }
 
-               my_assert(find_gene_id, line, "Cannot find gene_id!");
-               my_assert(find_transcript_id, line, "Cannot find transcript_id!");
+               my_assert(feature != "exon" || find_gene_id, line, "Cannot find gene_id!");
+               my_assert(feature != "exon" || find_transcript_id, line, "Cannot find transcript_id!");
+               if (!find_gene_id && feature != "exon") { printf("Warning: line \" %s \" does not contain a gene_id attribute! Since this line will not be used for reference construction, it is skipped. But if you think this GTF file is corrupted, you should find a complelete GTF file instead and rebuild the reference.\n", line.c_str()); }
+               if (!find_transcript_id && feature != "exon") { printf("Warning: line \" %s \" does not contain a transcript_id attribute! Since this line will not be used for reference construction, it is skipped. But if you think this GTF file is corrupted, you should find a complelete GTF file instead and rebuild the reference.\n", line.c_str()); }
        }
 
        std::string getSeqName() { return seqname; }

diff --git a/WHAT_IS_NEW b/WHAT_IS_NEW
index 2c6f2c7c33257c44434857a2359e855e44dd078e..da4c7641ec2eb721e6e975ab0b17d0dd198d683a 100644 (file)
--- a/WHAT_IS_NEW
+++ b/WHAT_IS_NEW
@@ -1,3 +1,9 @@
+RSEM v1.2.13
+
+- Changed the behavior of parsing GTF file. Now if a GTF line's feature is not "exon" and it does not contain a "gene_id" or "transcript_id" attribute, only a warning message will be produced (instead of failing the RSEM)
+
+--------------------------------------------------------------------------------------------
+
 RSEM v1.2.12
 
 - Enabled allele-specific expression estimation

diff --git a/samValidator.cpp b/samValidator.cpp
index 6e602107090f1d795341f9c08ee3b7d27689ae22..5586c8c3dbc1b6d46d8cd73b331fac976766296a 100644 (file)
--- a/samValidator.cpp
+++ b/samValidator.cpp
@@ -36,7 +36,7 @@ int main(int argc, char* argv[]) {
        size_t len = suffix.length();
        for (size_t i = 0; i < len; i++) suffix[i] = tolower(suffix[i]);
 
-       general_assert(suffix == "sam" || suffix == "bam", "Cannot recognize input file's file type! The file suffix is neither sam or bam.");
+       general_assert(suffix == "sam" || suffix == "bam", "Cannot recognize input file's file type! The file suffix is neither sam nor bam.");
 
        in = (suffix == "sam" ? samopen(argv[1], "r", NULL) : samopen(argv[1], "rb", NULL));
        general_assert(in != 0, "Cannot open input file!");

diff --git a/scanForPairedEndReads.cpp b/scanForPairedEndReads.cpp
index 7b20e11d9d9e81252f14f3f8c5984f96a005fe6d..3669f5faad46972a6b2949a2226a6a159f284b67 100644 (file)
--- a/scanForPairedEndReads.cpp
+++ b/scanForPairedEndReads.cpp
@@ -50,7 +50,7 @@ bool less_than(bam1_t *a, bam1_t *b) {
 
 int main(int argc, char* argv[]) {
        if (argc != 3) {
-               printf("UsaOAge: rsem-scan-for-paired-end-reads input.sam output.bam\n");
+               printf("Usage: rsem-scan-for-paired-end-reads input.sam output.bam\n");
                exit(-1);
        }