tophat running problem

[yasu]

Hi all,

I'm trying to run mRNA-seq for human by tophat (v1.0.12).
I succeeded to get proper output file in the preliminary dataset (first 100K reads from each .fq file). But I failed to get proper output in the real dataset (each contains ~17M reads).

I would appreciate any help you could give me with this.

Thanks in advance.

-Yasu

### preliminary_test ###

$ tophat -r 10 -p 8 -o tophat_hg19_test hg19 s_1_1.head4000000.fq,s_6_1.head4000000.fq,s_7_1.head4000000.fq s_1_2.head4000000.fq,s_6_2.head4000000.fq,s_7_2.head4000000.fq

[Mon Feb 8 13:31:03 2010] Beginning TopHat run (v1.0.12)
-----------------------------------------------
[Mon Feb 8 13:31:03 2010] Preparing output location tophat_hg19_test/
[Mon Feb 8 13:31:03 2010] Checking for Bowtie index files
[Mon Feb 8 13:31:03 2010] Checking for reference FASTA file
[Mon Feb 8 13:31:03 2010] Checking for Bowtie
Bowtie version: 0.11.3.0
[Mon Feb 8 13:31:03 2010] Checking reads
seed length: 43bp
format: fastq
quality scale: --phred33-quals
[Mon Feb 8 13:31:51 2010] Mapping reads against hg19 with Bowtie
[Mon Feb 8 13:34:24 2010] Joining segment hits
[Mon Feb 8 13:34:59 2010] Mapping reads against hg19 with Bowtie
[Mon Feb 8 13:37:30 2010] Joining segment hits
[Mon Feb 8 13:38:04 2010] Searching for junctions via segment mapping
[Mon Feb 8 13:44:59 2010] Retrieving sequences for splices
[Mon Feb 8 13:46:42 2010] Indexing splices
[Mon Feb 8 13:47:58 2010] Mapping reads against segment_juncs with Bowtie
[Mon Feb 8 13:48:47 2010] Joining segment hits
[Mon Feb 8 13:49:26 2010] Mapping reads against segment_juncs with Bowtie
[Mon Feb 8 13:50:15 2010] Joining segment hits
[Mon Feb 8 13:50:52 2010] Reporting output tracks
-----------------------------------------------
Run complete [00:33:58 elapsed]


### real_data ###

tophat -r 10 -p 8 -o tophat_hg19 hg19 s_1_1.fq,s_6_1.fq,s_7_1.fq s_1_2.fq,s_6_2.fq,s_7_2.fq

[Mon Feb 8 14:24:47 2010] Beginning TopHat run (v1.0.12)
-----------------------------------------------
[Mon Feb 8 14:24:47 2010] Preparing output location tophat_hg19/
[Mon Feb 8 14:24:47 2010] Checking for Bowtie index files
[Mon Feb 8 14:24:47 2010] Checking for reference FASTA file
[Mon Feb 8 14:24:47 2010] Checking for Bowtie
Bowtie version: 0.11.3.0
[Mon Feb 8 14:24:47 2010] Checking reads
seed length: 43bp
format: fastq
quality scale: --phred33-quals
[Mon Feb 8 14:39:23 2010] Mapping reads against hg19 with Bowtie
[Mon Feb 8 15:24:40 2010] Joining segment hits
[Mon Feb 8 15:35:38 2010] Mapping reads against hg19 with Bowtie
[Mon Feb 8 16:18:44 2010] Joining segment hits
[Mon Feb 8 16:18:44 2010] Searching for junctions via segment mapping
Warning: junction database is empty!
[Mon Feb 8 18:01:42 2010] Joining segment hits
[Mon Feb 8 18:11:38 2010] Joining segment hits
[Mon Feb 8 18:11:38 2010] Reporting output tracks
[FAILED]
Error: Report generation failed with err = 1
Traceback (most recent call last):
File "/bin/tophat", line 1518, in ?
sys.exit(main())
File "/bin/tophat", line 1490, in main
params.gff_annotation)
File "/bin/tophat", line 936, in compile_reports
exit(1)
TypeError: 'str' object is not callable

[yasu]

I add the report.log file from real_data (failed one).

### Real_data (reports.log) ###

tophat_reports v1.0.12
---------------------------------------
Error: cannot open map file for reading

#####################

Comparing with the run.log files from preliminary_test (succeeded one) and from real_data (failed one), "/bin/segment_juncs" doesn't work well.

Can somebody give me any help?

Thanks,

-Yasu

[Cole Trapnell]

The fact that TopHat thinks the seed length is 43bp is concerning. The default is 25, and it shouldn't be different unless you specified --segment-length, which you didn't. TopHat currently requires that FASTQ files have records where all of the nucleotides for each read appear on a single line. Same goes for the quality strings - all the quality characters need to be on one line. This is a limitation I haven't had time to fix yet. Can you verify that your FASTQ file is formatted this way?

[yasu]

Thanks for your kind help!!

My fastq file is something like this. I omitted the sequence+position id from the line after "+". Does this make the things bad?

-Yasu

###########
@HWI-EAS368:1:1:9:316#0/1
CTGGATGATAACATTCCAGAAGATGACTCAGGTGTCCCCACCC
+
BB66AB9ACBB@BCBAAA><BBBAAB?BBB@@@BA?B@BB@AB
@HWI-EAS368:1:1:9:424#0/1
CTCCCTGCCAGATATCGAGGAGGTGAAAGACCAGAGCAGGAAC
+
BCBBB>?>@CBABBB;A877??.:<<B@;@@?=>?A>?6AAA?
@HWI-EAS368:1:1:9:1060#0/1
TGGATGGTTCAGGATAATCACCTGAGCAGTGAAGCCAGCTGCT
+
BBBBB=?@BBB?=@A9AA@CAA><<@:5>7?=A?A@?=A???@
@HWI-EAS368:1:1:9:410#0/1
CGGAGGCGGAGGCTTGGGTGCGTTCAAGATTCAGCTTCACCCG
+
AA9AAA=:A7'=7=?4+366=AA@:A>999B:=2,=>1014>7
@HWI-EAS368:1:1:9:807#0/1
CGAACATTTCTGGCCCCCAAGTGTCAGCCCATTCACGTAAAAA
+
BBBBBBC@@<:;6BC>:@2<B=BBB@7;BB=:C@799:BBB?%
@HWI-EAS368:1:1:9:405#0/1
TGTAAAGCCTGAAACAGCTGCCTGTGTGGGACTGAGATGCAGG
+
?=>B=AA@AB@AAB?88>@@@BB>?B>A<=>?A81<-<@@B@@

[yasu]

I added the '--segment-length 25', but the comment is still as follows;

[Tue Feb 9 16:47:40 2010] Checking reads
seed length: 43bp
format: fastq
quality scale: --phred33-quals

Did I go some wrong way?

-Yasu