Tophat 2.0.4 and GTF conversion failure -- bug?

[mixter]

Hello,

I'm experiencing the following problem with tophat2 and haven't seen it described before nor able to debug it yet. I hope someone can make suggestions. Note that I can use my tophat installation normally when not supplying the option -G annotation.gtf

I ran tophat as follows:

$ tophat -G hg19.igenomeAnnotation2012.gtf -o TophatOut -p 4 --b2-sensitive genomes/hg19/hg19 Bcell.1.RNAseq.txt

[2012-07-09 15:17:50] Beginning TopHat run (v2.0.4)
-----------------------------------------------
[2012-07-09 15:17:50] Checking for Bowtie
                  Bowtie version:        2.0.0.6
[2012-07-09 15:17:50] Checking for Samtools
                Samtools version:        0.1.18.0
[2012-07-09 15:17:50] Checking for Bowtie index files
[2012-07-09 15:17:50] Checking for reference FASTA file
[2012-07-09 15:17:50] Generating SAM header for /genomes/hg19/hg19
        format:          fastq
        quality scale:   phred33 (default)
[2012-07-09 15:18:39] Reading known junctions from GTF file
        Warning: TopHat did not find any junctions in GTF file
[2012-07-09 15:18:39] Preparing reads
         left reads: min. length=36, max. length=36, 24797173 kept reads (41782 discarded)
[2012-07-09 15:26:55] Creating transcriptome data files..
[2012-07-09 15:27:50] Building Bowtie index from hg19.igenomeAnnotation2012.fa
        [FAILED]
Error: Couldn't build bowtie index with err = 1

I then looked at the converted GTF file, which was empty:

$ ls -la TophatOut/tmp
-rw-rw-r--. 1 me me       864  9. Jul 15:18 hg19_genome.bwt.samheader.sam
-rw-rw-r--. 1 me me         0  9. Jul 15:26 hg19.igenomeAnnotation2012.fa
-rw-rw-r--. 1 me me         0  9. Jul 15:18 hg19.juncs
-rw-rw-r--. 1 me me 817177304  9. Jul 15:26 left_kept_reads.bam
-rw-rw-r--. 1 me me    579376  9. Jul 15:26 left_kept_reads.bam.index
-rw-rw-r--. 1 me me       690  9. Jul 15:18 temp.samheader.sam

I then tried to convert the GTF to fasta manually, and it seems to just crash:

$ tophat/gtf_to_fasta genomes/hg19.igenomeAnnotation2012.gtf hg19.test.fa
tophat/gtf_to_fasta: /lib64/libz.so.1: no version information available (required by tophat/gtf_to_fasta)
terminate called after throwing an instance of 'std::logic_error'
  what():  basic_string::_S_construct NULL not valid

Due to the libz warning, I also tried compiling tophat from source (x64 linux system here, Scientific Linux 6.1 to be exact). This eliminated the warning, but I still got the error:

$ tophat-2.0.4/src/gtf_to_fasta genomes/hg19.igenomeAnnotation2012.gtf 
terminate called after throwing an instance of 'std::logic_error'
  what():  basic_string::_S_construct NULL not valid
Abgebrochen

Any suggestions would be much appreciated!

[mixter]

Addendum: This seems to be a non-issue. I found that the GTF file in question was actually corrupted.

Nicer error reporting to spot errors from tophat's utilities would be appreciated in future releases, though

[adamhfreedman]

I am getting the same error msg with gtf_to_fasta when run as a stand alone, i.e.

/lib64/libz.so.1: no version information available (required by /n/home05/afreedman/minreq/tophat-2.0.8.Linux_x86_64/gtf_to_fasta)
terminate called after throwing an instance of 'std::logic_error'
what(): basic_string::_S_construct NULL not valid



but when using the -G option with tophat, it runs just fine. In other words, the problem isn't with the format of the gtf file.

anyone have a sense of what might be going on and how to fix it?