TopHat: output is not recognized by UCSC

[bekkari]

Yep, the same I heard from author

UCSC recognizes nothing but chr#, so have to you build your own indexes on chromosome sequences, which can downloaded from UCSC.

[statsteam]

I had the same problem before. What you need to do is to make your own bowtie index file using 'bowtie-build' command on fa files you downloaded from UCSC site.

Hope it helps,
Statsteam

[geriatrics1200]

Hi, we ran TopHat without a GFF3 file and the output does not appear to be recognizable wiggle and bed formats. Here is what we got (first lines) of the wiggle file:
track type=bedGraph name="TopHat - read coverage"
gi|149258597|ref|NT_039457.7|Mm8_39497_37 0 46627 0
gi|149258597|ref|NT_039457.7|Mm8_39497_37 46627 46667 1
gi|149258597|ref|NT_039457.7|Mm8_39497_37 46667 48587 0
gi|149258597|ref|NT_039457.7|Mm8_39497_37 48587 48627 1

And here is what we got from the bed file:
track name=junctions description="TopHat junctions"
gi|149258597|ref|NT_039457.7|Mm8_39497_37 1132498 1133665 JUNC00000001
13 + 1132498 1133665 255,0,0 2 31,21 0,1146
gi|149258597|ref|NT_039457.7|Mm8_39497_37 1264693 1270688 JUNC00000002
39 + 1264693 1270688 255,0,0 2 32,32 0,5963

The program ran without errors. We think that there is a step missing that would associate the gi/NTs to chromosome coordinates. We are thus wondering what will be the way to work through this.. Is it because of the missing GFF3 file? (we could not find one for mouse, and if this is a problem, we would appreciate some pointers on how to make one) Any other suggestions?

Thanks very much in advance!