More errors with TopHat 1.1.0

[HTS]

Thanks Daehwan for quickly fixing the previous error of TopHat 1.1.0 on SE reads! As I am trying the updated version, I got stuck again at prep_reads with standard 75nt Illumina SE reads. The same command worked well under TopHat 1.0.14 so I think this must be a bug with the new version. The command line and errors are:

tophat -o tophat_cufflinks_test -a 6 -i 50 --solexa1.3-quals -F 0.05 -p 4 -g 20 --coverage-search --microexon-search -j ../../hg19.junc hg19 adipose_75nt.fq

[Mon Oct 4 22:33:39 2010] Beginning TopHat run (v1.1.0)
-----------------------------------------------
[Mon Oct 4 22:33:39 2010] Preparing output location tophat_cufflinks_test/
[Mon Oct 4 22:33:39 2010] Checking for Bowtie index files
[Mon Oct 4 22:33:39 2010] Checking for reference FASTA file
[Mon Oct 4 22:33:39 2010] Checking for Bowtie
Bowtie version: 0.12.7.0
[Mon Oct 4 22:33:39 2010] Checking for Samtools
Samtools version: 0.1.8.0
[Mon Oct 4 22:33:40 2010] Checking reads
min read length: 75bp, max read length: 75bp
format: fastq
quality scale: phred64 (reads generated with GA pipeline version >= 1.3)
[FAILED]
Error: could not execute prep_reads

Anybody has the same problem or can offer a fix? Thanks a lot!

-- Leo

[kopi-o]

I have the same problem as well, "could not execute prep_reads".

[Telor]

I had this problem also, it was because the executable "prep_reads" provided with the pre-built linux distribution was crashing with a floating point exception. Once I built the code from the source distribution it worked fine.
(this problem with floating point exceptions is something I have always seen with the cufflinks linux distribution on my system, so I have had to resort to building the executables from source for all these - at least building tophat is much more straight forward! )

I hope this helps.

[Cole Trapnell]

Hm - I wonder if something in the build scripts for TopHat and Cufflinks is causing the compiler to generate code that incompatible with your architecture. Can you tell me a little about the machines you guys are seeing this crash on? I'll try to fix the scripts so that the binaries are portable.

[kopi-o]

I'm running it on a Mac OS X 10.6.4 (Intel Core i7 64-bit)

[HTS]

I am running Ubuntu 10.04 LTS 64-bit

[Telor]

Hi Cole,

thanks for the reply,

This is an issue I've had with cufflinks since version 0.8.2. I was running 0.8.1 until now because I wasn;t able to get the executalbes to work. I only now have had some success with compiling the software on my system (with your new helpful instructions about the Bam libraries), but unfortunately I'm still getting crashes when it gets to the "Assembling transcripts and estimating abundances" stage (maybe its to do with problems with libaries etc..).

That said, usually the Tophat and bowtie executables work fine (except for the latest release with "prep_reads")

My system is:
SUSE Linux Enterprise Server 10 SP2 (x86_64) GNU/Linux

It would be great if it was possible to get a working executable.

Thanks,

Chris.

[krobison]

I saw this problem with my SE colorspace reads from SRA (after hacking the tophat code to get past the length(quality)=length(read) issue with the SRA file).

I'm on 64-bit Enterprise Oracle Linus, which is in the RedHat family of Linuxes,

[hyjkim]

I am getting the same read_prep error. Here is the command I am issuing:

 tophat -C -Q --integer-quals -o tophat-out -p 4 mm9_c test.csfasta test.qual

tophat-out/logs/prep_reads.log

prep_reads v1.1.0 (1606)
---------------------------
Saw a space but expected an ASCII-encoded quality value.
Are quality values formatted as integers?  If so, try --integer-quals.
terminate called after throwing an instance of 'int'

If I try to run tophat without a quality file using:

  tophat -C -o tophat-out -p 4 mm9_c test.csfasta

I receive the following error:

Traceback (most recent call last):
  File "/share/apps/tophat-1.1.0.Linux_x86_64/tophat", line 2174, in ?
    sys.exit(main())
  File "/share/apps/tophat-1.1.0.Linux_x86_64/tophat", line 2067, in main
    left_quals_list = args[2]
IndexError: list index out of range

[telos]

I'm stuck with the same prep_reads problem as hyjkim when using csfasta+qual files.

One other issue: the new tophat doesn't cope with csfasta/qual files that have #-comments at the top.

[dcjones]

The prep_reads problem seems to be caused by -1 qualities.

Here is a simple fix (replacing -1 qualities with 0):

--- src/reads.cpp
+++ src/reads.cpp
@@ -119,7 +119,7 @@ bool next_fastq_record(FLineReader& fr,
        string temp_qual;
        for (size_t i = 0; i < integer_qual_values.size(); ++i)
          {
-           temp_qual.push_back((char)(atoi(integer_qual_values[i].c_str()) + 33));
+           temp_qual.push_back((char)(max(0,atoi(integer_qual_values[i].c_str())) + 33));
          }
 
        qual.append(temp_qual);

[hyjkim]

The prep_reads problem seems to be caused by -1 qualities.

Here is a simple fix (replacing -1 qualities with 0):

--- src/reads.cpp
+++ src/reads.cpp
@@ -119,7 +119,7 @@ bool next_fastq_record(FLineReader& fr,
        string temp_qual;
        for (size_t i = 0; i < integer_qual_values.size(); ++i)
          {
-           temp_qual.push_back((char)(atoi(integer_qual_values[i].c_str()) + 33));
+           temp_qual.push_back((char)(max(0,atoi(integer_qual_values[i].c_str())) + 33));
          }
 
        qual.append(temp_qual);

Thanks dcjones. I replaced all the -1 qualities in my .qual file with 0's. Tophat is no longer crashing as it was before. I'll report once the mapping run is complete.

[twaddlac]

Hello,

I am receiving a similar error to hyjkim when trying to run tophat using the command

tophat --solexa-quals ref_seq data.fastq

Here's the error from stdout:

[Thu Mar 10 13:35:19 2011] Beginning TopHat run (v1.2.0)
-----------------------------------------------
[Thu Mar 10 13:35:19 2011] Preparing output location ./tophat_out/
[Thu Mar 10 13:35:19 2011] Checking for Bowtie index files
[Thu Mar 10 13:35:19 2011] Checking for reference FASTA file
[Thu Mar 10 13:35:19 2011] Checking for Bowtie
	Bowtie version:			 0.12.7.0
[Thu Mar 10 13:35:19 2011] Checking for Samtools
	Samtools Version: 0.1.7
[Thu Mar 10 13:35:19 2011] Checking reads
	min read length: 54bp, max read length: 54bp
	format:		 fastq
	quality scale:	 phred33 (default)
[Thu Mar 10 13:38:43 2011] Mapping reads against bm-ref-seq-cut with Bowtie
[Thu Mar 10 13:38:43 2011] Joining segment hits
[Thu Mar 10 13:38:44 2011] Searching for junctions via segment mapping
Traceback (most recent call last):
  File "/usr/local/bin/tophat", line 2346, in <module>
    sys.exit(main())
  File "/usr/local/bin/tophat", line 2304, in main
    user_supplied_deletions)
  File "/usr/local/bin/tophat", line 2030, in spliced_alignment
    ref_fasta)
  File "/usr/local/bin/tophat", line 1736, in junctions_from_segments
    slash = left_seg_maps[0].rfind('/')
IndexError: list index out of range

and here's the log file:

-p/--threads is disabled because bowtie was not compiled with pthreads support
Command: bowtie -q --un ./tophat_out/tmp/left_kept_reads_missing.fq --max /dev/null -v 2 -p 1 -k 40 -m 40 bm-ref-seq-cut ./tophat_out/tmp/left_kept_reads.fq

I know that I have the pthreads library but I'm thinking it's more of an input error? Any ideas?

Thank you!

[mart555]

The prep_reads problem seems to be caused by -1 qualities.

Here is a simple fix (replacing -1 qualities with 0):

--- src/reads.cpp
+++ src/reads.cpp
@@ -119,7 +119,7 @@ bool next_fastq_record(FLineReader& fr,
        string temp_qual;
        for (size_t i = 0; i < integer_qual_values.size(); ++i)
          {
-           temp_qual.push_back((char)(atoi(integer_qual_values[i].c_str()) + 33));
+           temp_qual.push_back((char)(max(0,atoi(integer_qual_values[i].c_str())) + 33));
          }
 
        qual.append(temp_qual);

I also got this problem, but I don't know how to use your script, can you tell me more about how to replace -1 with 0 in .qual file?