Tophat 2.0.3: Warnings

[Droplet]

Hi all,

I am completely new to tophat. It seems that it did not run properly as it generated two warning messages and all bed files were empty (including deletions, insertions and junctions). In addition, the size of the accepted_hit.bam file is only 250kb while the unmapped.bam is 1.1GB. For details, please see the code below.

It would be very appreciated if anyone could help me solve these issues:
1) Why is the junction database empty? (Waring1)
2) How to remove all temporary files? (Warning2)
3) Why are the bed files empty?
4) Why are there so many reads unmapped?

Thanks!

tophat --bowtie1 --color --segment-length 13 -o /home/usr/Download/Data2 hg18 /home/usr/Download/Sample_R3.csfasta.txt
[2012-06-11 19:33:18] Beginning TopHat run (v2.0.3)
-----------------------------------------------
[2012-06-11 19:33:18] Checking for Bowtie
                  Bowtie version:        0.12.8.0
[2012-06-11 19:33:18] Checking for Samtools
                Samtools version:        0.1.18.0
[2012-06-11 19:33:18] Checking for Bowtie index files
[2012-06-11 19:33:18] Checking for reference FASTA file
[2012-06-11 19:33:18] Generating SAM header for hg18
        format:          fasta
[2012-06-11 19:33:18] Preparing reads
         left reads: min. length=25, max. length=25, 81854203 kept reads (0 discarded)
[2012-06-11 19:49:15] Mapping left_kept_reads to genome hg18 with Bowtie 
[COLOR="Red"]Warning: junction database is empty![/COLOR]
[2012-06-11 19:51:52] Reporting output tracks
[COLOR="Red"]Warning: couldn't remove all temporary files in /home/usr/Download/Data2/tmp/[/COLOR]
-----------------------------------------------
[2012-06-11 20:21:16] Run complete: 00:47:58 elapsed

[Dario1984]

Hello,

I have the same problem. My command is :

tophat -p 4 --genome-read-mismatches 3 --read-mismatches 3 --max-multihits 1 --b2-sensitive hg19 /home/darstr/tmp/ArrayStarProbes.fa

The sequences in the FASTA are microarray probes either 60 or 61 bases long, some designed across splice junctions. I'm mapping them with TopHat to re-annotate them.

$ head /home/darstr/tmp/ArrayStarProbes.fa
>CUST_385_PI426073487
ACTTGTTTCAGGTTAATGGTAAGAGTTACCCGCAGGCTAAGTTGCTATTGGGACAGATGG
>ASHG19A3A032313
GAAACATCTTCCCAGAAAACAGACTACAACAAATCATTTCCCTGGTATGAAGGAATGAAG
>ASHG19A3L0000337
AGATATCTGGTTGAGACCAAGTAAGAGATAAACCTGTAGATTCCACTGTGCCACCACAGT
>ASHG19A3A014810
GGACAAGAATCCTTCAAAAAACAGGAAAAAACTCCTAAAACACCTAAAAGGACCTAGTTC
>ASHG19A3A015967
AAGATCTTCATAGAAAGAGTCCAGGGGAACCCAACCCTATTGCACATCACAGATCTTCAG

I get 0 successfully mapped sequences.

$ samtools view -c tophat_out/accepted_hits.bam
0

I know there's quite a few uniquely mapping perfectly matching probes on the array.

I got the same 2 warning messages from my run as the original poster. I hope a new TopHat version is released soon.