a question about tophat

[lei]

I am using tophap with solid single-end data.
the command line is :
tophat --color --quals -p 8 -o $outdir --library-type fr-secondstrand -G $gtffile hg19 $reads $qual

it reports the error message:
Error: you must set the mean inner distance between mates with -r

but -r is only for paired-end data, so I don't know what should I do.

[seqguy]

Dear Lie

You have to rearrange tophat parameters to run it successfully. Tyr the below mentioned command:

./tophat -G <path to .gtf file> --color --quals <path to index folder/index prefix> <path to .csfasta file> <path to .qual file>

[lei]

Dear seqguy
Thanks a lot!
I runned the command in your order. It started to run, but reported another error:
IOError: [Errno 2] No such file or directory: '/home/lei/solid/mec1_f3/tmp/left_kept_reads_missing.fq'
But I can find the file left_kept_reads_missing.fq in the tmp directory.
So do you have any idea what's going on this time?

[seqguy]

Could you please let me know the configuration of the system you are using to run this pipeline, the organism genome size and also if you are using root user or not.

[lei]

it was running on a Server, the CPU is 24x 2.3 GHz, the Memory is 80 GB, and the OS is suse11.0.
I used the human genome from UCSC.
I'm not the root user.

[lei]

It's strange that this time, I can run it with my previous parameters' order. But still report the same error message: IOError: [Errno 2] No such file or directory: '/home/lei/solid/mec1_f3/tmp/left_kept_reads_missing.fq'

[seqguy]

Have you created the bowtie index of your reference genome with -C option?

[lei]

sure, I runned tophat before with illumina data, everything was fine. those error just happened when I use solid data.

[seqguy]

1) Check if your GTF and reference genome files are having same naming convention for chromosomes.
2) Reference genome index should be in colorspace.
3) Check if .csfasta and .qual files are having same counts.

[lei]

yes, 1),2),3) are all fine.
1) it's possible that the genome fasta file has "chr1" etc. whereas the GTF file has "1" etc. but I ran cuffdiff from the same author successfully with the same GTF file and the original genome from which the bowtie index was built is the same as the one I used there.
3) the alignments I ran would never have worked if there had been a different number of reads in the csfasta and .qual files.

Maybe I need to pose my question again to make it clear:

The first time I run the tophat using qsub. It reports the error:
Error: you must set the mean inner distance between mates with -r

Then, I rearrange the parameters, test it in a local machine with both parameters' orders mentioned above, get another error:
IOError: [Errno 2] No such file or directory: '/home/lei/solid/mec1_f3/tmp/left_kept_reads_missing.fq'

Now, I test it using qsub again, get the same error again:
Error: you must set the mean inner distance between mates with -r

[seqguy]

could you please share the exact command you are using to run tophat?

[lei]

the first script is :

#!/bin/bash

#PBS -N Tophat_lib1
#PBS -l nodes=1pn=8:lsdf
#PBS -l walltime=50:00:00
#PBS -l mem=5g
#PBS -M [email protected]
#PBS -m ae
#PBS -j eo
#PBS -e /home/gu/solid/lib1/error

# environment variable for the indexes
BOWTIE_INDEXES=/home/gu/solid/bowtieIndex/
export BOWTIE_INDEXES
reads=/home/gu/solid/lib1/138000338_20101201_FRAG_BC_SEQ0014_bcSample1_lib1_F3.csfasta
qual=/home/gu/solid/lib1/138000338_20101201_FRAG_BC_SEQ0014_bcSample1_lib1_F3_QV.qual
outdir= /home/gu/solid/lib1
gtffile=/ibios/co01/hutter/PlassData/Ensembl57_transcripts_hg19.gtf

tophat -G $gtffile --color --quals -p 8 -o $outdir --library-type fr-secondstrand hg19 $reads $qual

the error code is :Error: you must set the mean inner distance between mates with -r

[lei]

the econd script is :

#!/bin/bash

#PBS -N Tophat_lib1
#PBS -l nodes=1pn=8:lsdf
#PBS -l walltime=50:00:00
#PBS -l mem=5g
#PBS -M [email protected]
#PBS -m ae
#PBS -j eo
#PBS -e /home/gu/solid/lib1/error

# environment variable for the indexes
BOWTIE_INDEXES=/home/gu/solid/bowtieIndex/
export BOWTIE_INDEXES
#reads=/home/gu/solid/lib1/138000338_20101201_FRAG_BC_SEQ0014_bcSample1_lib1_F3.csfasta
#qual=/home/gu/solid/lib1/138000338_20101201_FRAG_BC_SEQ0014_bcSample1_lib1_F3_QV.qual
#outdir= /home/gu/solid/lib1
#gtffile=/ibios/co01/hutter/PlassData/Ensembl57_transcripts_hg19.gtf

tophat --color --quals -p 8 -o /home/gu/solid/lib1 --library-type fr-secondstrand -G /ibios/co01/hutter/PlassData/Ensembl57_transcripts_hg19.gtf /icgc/lsdf/mb/analysis/hutter/bowtieIndex/hg19 /home/gu/solid/lib1/138000338_20101201_FRAG_BC_SEQ0014_bcSample1_lib1_F3.csfasta /home/gu/solid/lib1/138000338_20101201_FRAG_BC_SEQ0014_bcSample1_lib1_F3_QV.qual

the error code is :
[Thu Mar 17 16:12:28 2011] Beginning TopHat run (v1.2.0)
-----------------------------------------------
[Thu Mar 17 16:12:28 2011] Preparing output location /home/gu/solid/lib1/
[Thu Mar 17 16:12:28 2011] Checking for Bowtie index files
[Thu Mar 17 16:12:28 2011] Checking for reference FASTA file
[Thu Mar 17 16:12:28 2011] Checking for Bowtie
Bowtie version: 0.12.7.0
[Thu Mar 17 16:12:28 2011] Checking for Samtools
Samtools Version: 0.1.12a
[Thu Mar 17 16:12:28 2011] Checking reads
min read length: 50bp, max read length: 50bp
format: fasta
[Thu Mar 17 16:14:07 2011] Reading known junctions from GTF file
[Thu Mar 17 16:18:51 2011] Mapping reads against hg19 with Bowtie
[Thu Mar 17 16:18:51 2011] Joining segment hits
Traceback (most recent call last):
File "/ibios/tbi_cluster/11.0/x86_64/bin/tophat", line 2346, in <module>
sys.exit(main())
File "/ibios/tbi_cluster/11.0/x86_64/bin/tophat", line 2304, in main
user_supplied_deletions)
File "/ibios/tbi_cluster/11.0/x86_64/bin/tophat", line 1972, in spliced_alignment
segment_len)
File "/ibios/tbi_cluster/11.0/x86_64/bin/tophat", line 1545, in split_reads
reads_file = open(reads_filename)
IOError: [Errno 2] No such file or directory: '/home/gu/solid/lib1/tmp/left_kept_reads_missing.fq'