Tophat can not find bowtie?

[qiudao]

Dear List,

I tried to use tophat for analysis of rna-seq.

However, when after I run tophat test data, it gave me the error:

./tophat -r 20 test_ref reads_1.fq reads_2.fq

Beginning TopHat run (v1.2.0)
Preparing output location ./tophat_out/
Checking for Bowtie index files
Checking for reference FASTA file
Checking for Bowtie
Error: TopHat requires Bowtie 0.12.3 or later

But I have already put the directory of "bowtie" which contains "bowtie","bowtie-build" and "bowtie-inspect" in $PATH and export it when starting the shell.

I tried two version of tophat, x86 binary and my own compiled, they gave me the same errors; I also tried two versions bowtie (0.12.5 and 0.12.7), both gave me the same error.

could someone tell me what I was missing? Thanks.

-Q

[qiudao]

Dear List,

I have tried to copy these binaries to another 64bit machine:
samtools, bowtie 64bit binary, and tophat 64bit binary. And chang the PATH to have bowtie and samtools
It works fine.

However, the same setting for my previous machine (a cluster's front end), it dosen't work, and gave me the error again:
Beginning TopHat run (v1.2.0)
Preparing output location ./tophat_out/
Checking for Bowtie index files
Checking for reference FASTA file
Checking for Bowtie
Error: TopHat requires Bowtie 0.12.3 or later

Does anyone have any idea about this?
Thank you.
-Q

[NM_010117]

You could try altering your .bashrc file again and exporting BOWTIE_HOME since exporting the appended PATH variable doesn't seem to be working.

e.g. part of your .bashrc would look like this:

BOWTIE_HOME=~/bin/bowtie/
export BOWTIE_HOME

[qiudao]

Hi NM_010117,

Thank you for your reply. I just realized that someone installed an old version bowtie on our system which supersedes my new version. I have changed setting to make it work.
So it's nothing wrong with tophat or bowtie.
Thanks.

-Q

[mattarno]

similar bowtie version issues

hi - i am having a similar problem with versions of bowtie/bowtie2 on a server. you mentioned:

"I have changed setting to make it work."

it would be great if you could expand on that...

basically as far as i can tell my server has multiple versions of bowtie including the up-to-date version installed and in correctly in the path, but when I run tophat, it errors and says

[2015-05-21 14:37:29] Checking for Bowtie
Bowtie 2 not found, checking for older version..
Error: TopHat requires Bowtie 0.12.9 or later

if i run bowtie --version i get 0.12.7

and bowtie2 --version i get permission denied...


cheers for any input

Matt

[GenoMax]

Are you able to work with your administrators on this since that would be the easiest way to fix this problem (unless you are the administrator).

bowtie v.1 - Default appears to be set to 0.12.7. If that is the latest available then upgrade to the latest version which is 1.1.1.

As for bowtie2 - add execute permissions to all bowtie2* programs. Bowtie2 is currently in v.2.2.5.

In general for Tuxedo suite programs it is best to upgrade to latest software versions available to avoid problems.

[mattarno]

great - thanks for quick reply - i am not an admin (and a fairly novice unix user, too). i've checked permissions on bowtie2 it says '-rwxrwx---' which i guess means user has no rights --- (last three chrs)

i will contact admin

cheers,
matt

[mattarno]

ok - permissions now sorted. So i tried again. I am following tophat manual for making initial tophat/bowtie indexes from my GTF file:

"For example, in order to just prepare the transcriptome index files for a specific annotation, an initial, single TopHat run could be invoked like this:

tophat -G known_genes.gtf \
--transcriptome-index=transcriptome_data/known \
hg19 "

when i set this up in the dir with the gtf file I get the following output:

$ tophat2 -G hg19_genes.gtf --transcriptome-index=transcriptome_data/known hg19

[2015-05-21 16:44:12] Building transcriptome files with TopHat v2.0.13
-----------------------------------------------
[2015-05-21 16:44:12] Checking for Bowtie
Bowtie version: 2.2.5.0
[2015-05-21 16:44:15] Checking for Bowtie index files (genome)..
Error: Could not find Bowtie 2 index files (hg19.*.bt2)


what am I doing wrong? I thought this command would make tyhe index files, so why is it looking for them first? Do i need more than just my GTF file? Have I got the wrong end of the stick?

CHeers,
matt

[GenoMax]

ok - permissions now sorted. So i tried again. I am following tophat manual for making initial tophat/bowtie indexes from my GTF file:

"For example, in order to just prepare the transcriptome index files for a specific annotation, an initial, single TopHat run could be invoked like this:

tophat -G known_genes.gtf \
--transcriptome-index=transcriptome_data/known \
hg19 "

when i set this up in the dir with the gtf file I get the following output:

$ tophat2 -G hg19_genes.gtf --transcriptome-index=transcriptome_data/known hg19

[2015-05-21 16:44:12] Building transcriptome files with TopHat v2.0.13
-----------------------------------------------
[2015-05-21 16:44:12] Checking for Bowtie
Bowtie version: 2.2.5.0
[2015-05-21 16:44:15] Checking for Bowtie index files (genome)..
Error: Could not find Bowtie 2 index files (hg19.*.bt2)


what am I doing wrong? I thought this command would make tyhe index files, so why is it looking for them first? Do i need more than just my GTF file? Have I got the wrong end of the stick?

CHeers,
matt

So the idea here is that you are creating a subset (indexes) that covers just the transcriptome from the entire genome. For that to work you do need to start with bowtie2indexes for the whole genome.

You can get that annotation/sequence/index bundle from the iGenomes site here: http://support.illumina.com/sequenci...e/igenome.html Pick the build that you like (UCSC, NCBI etc).

[mattarno]

thanks - can you point me to a resource which guides me through that part, i.e. how to incorporate these files into the tophat transcriptome index process? I'm a bit lost at the moment, and I have plenty of time while I download 25GB of iGenome files..

Cheers,
matt

[GenoMax]

Your command will look something like this (using the iGenomes files for annotation and index). Replace "path_to" with real paths on your system.

$ tophat2 -G /path_to/Annotations/Genes/genes.gtf --transcriptome-index=transcriptome_data/known /path_to/Bowtie2Index/genome

In your susequent tophat runs you will be referring to the transcriptome index

$ tophat2 --library-type XXXX -o sample_ID_th_out \
-T --transcriptome-index=/path_to/transcriptome_data/known \
/path_to/Sequence/Bowtie2Index/genome sample_ID_R1.gz sample_ID_fastq_R2.gz

[mattarno]

that's great - thanks - it's starting to come together for me now

cheers,
matt

[mattarno]

hello - I'm stuck again. I downloaded and untarred the iGenome file. trying with the above script - added in the paths, but the script trips up on the transcriptome_data/known and the bowtie2Index lines:

the script is like this:

tophat2 -G /home/marnobrc/genome/Homo_sapiens/UCSC/hg19/Annotation/Archives/archive-2014-06-02-13-47-56/Genes/genes.gtf
--transcriptome-index=/home/marnobrc/genome/Homo_sapiens/UCSC/hg19/transcriptome_data/known
/home/marnobrc/genome/Homo_sapiens/UCSC/hg19/Sequence/Bowtie2Index/genome

but it comes back with these errors:

/usr/local/sge/slam/spool/node13/job_scripts/1639683: line 15: --transcriptome-index=/home/marnobrc/genome/Homo_sapiens/UCSC/hg19/transcriptome_data/known: No such file or directory
/usr/local/sge/slam/spool/node13/job_scripts/1639683: line 16: /home/marnobrc/genome/Homo_sapiens/UCSC/hg19/Sequence/Bowtie2Index/genome: No such file or directory

so i've tried creating the transcriptome_data dir in hg19 but it makes no difference...

here's the contents of the hg19 dir (with all the stuff from the iGenomes file installed)...

$ pwd
/home/marnobrc/genome/Homo_sapiens/UCSC/hg19
$ ls -l
total 32625288
drwxrwxr-x 3 marnobrc marnobrc        4096 May 21 16:40 annotation
drwxrwxr-x 3 marnobrc marnobrc        4096 Jun  2  2014 Annotation
-rw-rw-r-- 1 marnobrc marnobrc 29666884240 May 22 06:24 Homo_sapiens_UCSC_hg19.tar.gz
drwxrwxr-x 9 marnobrc marnobrc        4096 May 28  2014 Sequence
drwxrwxr-x 2 marnobrc marnobrc        4096 May 23 00:09 transcriptome_data
-rw-r--r-- 1 marnobrc marnobrc       16206 May 22 13:33 untar_igenome.o1639618
-rw-rw-r-- 1 marnobrc marnobrc         152 May 22 13:10 untar_igenome.sh

and here's the Sequence dir with the bowtie2Indexes

$ pwd
/home/marnobrc/genome/Homo_sapiens/UCSC/hg19/Sequence
$ ls -l
total 112
drwxrwxr-x 2 marnobrc marnobrc 4096 Jun  3  2014 AbundantSequences
drwxrwxr-x 2 marnobrc marnobrc 4096 Jun 13  2013 BlastDB
drwxrwxr-x 2 marnobrc marnobrc 4096 Apr 11  2012 Bowtie2Index
drwxrwxr-x 2 marnobrc marnobrc 4096 Apr 11  2012 BowtieIndex
drwxrwxr-x 4 marnobrc marnobrc 4096 Apr 27  2012 BWAIndex
drwxrwxr-x 2 marnobrc marnobrc 4096 Mar 16  2012 Chromosomes
drwxrwxr-x 2 marnobrc marnobrc 4096 May 19  2014 WholeGenomeFasta

and the contents of bowtie2Index dir...

$ pwd
/home/marnobrc/genome/Homo_sapiens/UCSC/hg19/Sequence/Bowtie2Index
$ ls -l
total 4467512
-rwxr-xr-x 1 marnobrc marnobrc 957980027 Apr 10  2012 genome.1.bt2
-rwxr-xr-x 1 marnobrc marnobrc 715335932 Apr 10  2012 genome.2.bt2
-rwxr-xr-x 1 marnobrc marnobrc      3284 Apr 10  2012 genome.3.bt2
-rwxr-xr-x 1 marnobrc marnobrc 715335926 Apr 10  2012 genome.4.bt2
lrwxrwxrwx 1 marnobrc marnobrc        29 May 22 13:33 genome.fa -> ../WholeGenomeFasta/genome.fa
-rwxr-xr-x 1 marnobrc marnobrc 957980027 Apr 10  2012 genome.rev.1.bt2
-rwxr-xr-x 1 marnobrc marnobrc 715335932 Apr 10  2012 genome.rev.2.bt2

please help. where have I gone wrong? Have i missed something?

Matt

[mattarno]

wait...did I just not put the "\" backslashes in the script at the ends of lines???

I just put them in and it's doing something different...

[2015-05-23 00:37:27] Building transcriptome files with TopHat v2.0.13
-----------------------------------------------
[2015-05-23 00:37:27] Checking for Bowtie
                  Bowtie version:        2.2.5.0
[2015-05-23 00:37:29] Checking for Bowtie index files (genome)..
[2015-05-23 00:37:29] Checking for reference FASTA file
[2015-05-23 00:37:29] Building transcriptome data files /home/marnobrc/genome/Homo_sapiens/UCSC/hg19/transcriptome_data/known
[2015-05-23 00:38:00] Building Bowtie index from known.fa

i hope that was all it was

..waiting with baited breath..

matt

[GenoMax]

The "\" just splits the command line so it becomes readable. I think you are fine based on the snippet above.