warning message on junction databse

[aghazalp]

Hello everyone, I just signed up and wanted to say hi and also ask a question. I just installed tophat and bowtie and created a test file from my large fastq file. This test file has only ten sequences in it (40 lines total). When I run the tophat command it took a very short time and produced the following verbose and an output directory as was supposed to. My question is, in the verbose below there is a statement saying "Warning: junction database is empty!". Can anyone kindly explain what this means or perhaps refer me to a source where I could read about it more. The verbose is pasted below. Thank you for your time.

[Wed Jun 2 12:10:20 2010] Beginning TopHat run (v1.0.13)
-----------------------------------------------
[Wed Jun 2 12:10:20 2010] Preparing output location ./tophat_out/
[Wed Jun 2 12:10:20 2010] Checking for Bowtie index files
[Wed Jun 2 12:10:20 2010] Checking for reference FASTA file
[Wed Jun 2 12:10:20 2010] Checking for Bowtie
Bowtie version: 0.12.5.0
[Wed Jun 2 12:10:20 2010] Checking reads
seed length: 76bp
format: fastq
quality scale: phred33 (default)
[Wed Jun 2 12:10:20 2010] Mapping reads against mm9 with Bowtie
[Wed Jun 2 12:10:53 2010] Joining segment hits
Splitting reads into 3 segments
[Wed Jun 2 12:10:53 2010] Mapping reads against mm9 with Bowtie
[Wed Jun 2 12:11:28 2010] Mapping reads against mm9 with Bowtie
[Wed Jun 2 12:12:02 2010] Mapping reads against mm9 with Bowtie
[Wed Jun 2 12:12:36 2010] Searching for junctions via segment mapping
Warning: junction database is empty!
[Wed Jun 2 12:12:36 2010] Joining segment hits
[Wed Jun 2 12:12:37 2010] Reporting output tracks
-----------------------------------------------
Run complete [00:02:17 elapsed]

[dariober]

My question is, in the verbose below there is a statement saying "Warning: junction database is empty!". Can anyone kindly explain what this means or perhaps refer me to a source where I could read about it more. The verbose is pasted below. Thank you for your time.

[Wed Jun 2 12:10:20 2010] Beginning TopHat run (v1.0.13)
-----------------------------------------------
[Wed Jun 2 12:10:20 2010] Preparing output location ./tophat_out/
[Wed Jun 2 12:10:20 2010] Checking for Bowtie index files
[Wed Jun 2 12:10:20 2010] Checking for reference FASTA file
[Wed Jun 2 12:10:20 2010] Checking for Bowtie
Bowtie version: 0.12.5.0
[Wed Jun 2 12:10:20 2010] Checking reads
seed length: 76bp
format: fastq
quality scale: phred33 (default)
[Wed Jun 2 12:10:20 2010] Mapping reads against mm9 with Bowtie
[Wed Jun 2 12:10:53 2010] Joining segment hits
Splitting reads into 3 segments
[Wed Jun 2 12:10:53 2010] Mapping reads against mm9 with Bowtie
[Wed Jun 2 12:11:28 2010] Mapping reads against mm9 with Bowtie
[Wed Jun 2 12:12:02 2010] Mapping reads against mm9 with Bowtie
[Wed Jun 2 12:12:36 2010] Searching for junctions via segment mapping
Warning: junction database is empty!
[Wed Jun 2 12:12:36 2010] Joining segment hits
[Wed Jun 2 12:12:37 2010] Reporting output tracks
-----------------------------------------------
Run complete [00:02:17 elapsed]

I guess there is nothing strange here. With only 10 reads is not unlikely that there is no junction being detected (unless you selected the 10 reads on purpose to span some junctions).

Dario

[aghazalp]

Hi Dario,

I was thinking the same way as you mentioned in your reply so I just run the entire sequence file which took me about 8 hours for about 20 million sequences but I got the same verbose which I am pasting below. So there is definitely something not right. I am using the UCSC mm9 genome downloaded from the tophat website if that matters. here are my commands and the verbose of the output:

tophat -o AG_tophat_result_reads1 /Users/anatoleghazalpour/ExeFiles/bowtie-0.12.5/indexes/mm9 reads.1.fastq

[Wed Jun 2 14:52:16 2010] Preparing output location AG_tophat_result_reads1/
[Wed Jun 2 14:52:16 2010] Checking for Bowtie index files
[Wed Jun 2 14:52:16 2010] Checking for reference FASTA file
[Wed Jun 2 14:52:16 2010] Checking for Bowtie
Bowtie version: 0.12.5.0
[Wed Jun 2 14:52:16 2010] Checking reads
seed length: 76bp
format: fastq
quality scale: phred33 (default)
[Wed Jun 2 15:02:05 2010] Mapping reads against mm9 with Bowtie
[Wed Jun 2 16:03:26 2010] Joining segment hits
Splitting reads into 3 segments
[Wed Jun 2 16:10:15 2010] Mapping reads against mm9 with Bowtie
[Wed Jun 2 18:21:45 2010] Mapping reads against mm9 with Bowtie
[Wed Jun 2 21:43:26 2010] Mapping reads against mm9 with Bowtie
[Wed Jun 2 23:22:15 2010] Searching for junctions via segment mapping
Warning: junction database is empty!
[Wed Jun 2 23:25:35 2010] Joining segment hits
[Wed Jun 2 23:32:40 2010] Reporting output tracks

[dariober]

Hello,
I can't tell much... but in case you haven't already done so, have a look at the log files in log directory. Is there anything suspicious? How many reads have been aligned?

Good luck!
Dario

[aghazalp]

thanx dario, I looked at them and nothing stands out. Am I the only one who has this problem?

What can help me is if I could get my hands on a set of reads from mouse RNA which I know should work to test if it is my fastq file is correctly formatted or it is the tophat that is not doing what it is supposed to. Any idea where I can get my hands on a fastq file which has been tested before?

thanx,

Anatole