cufflinks transcript on wrong strand

[ornam]

Hi,

I've run tophat (v.1.1.4) and cufflinks (v0.9.2) on paired-end stranded RNA-seq data.
When I examined the transcripts produced by cufflinks for my favorite gene (using the UCSC genome browser), I found to my dismay that the transcript was on the wrong strand. When I examined the junctions.bed file for the region the junctions are on the correct strand.
Below are the commands I used for tophat and cufflinks:
tophat -p 12 -o SC3_s_7_tophat_defaults_out -r 48 --mate-std-dev 27 --solexa1.3-quals --coverage-search --microexon-search --library-type fr-unstranded hg19 SC3_s_7_1_sequence.txt SC3_s_7_2_sequence.txt

cufflinks accepted_hits.bam -L SC3 -p 20 -r /lab/db/bowtie/hg19.fa --library-type fr-firststrand

Any insight/help would be appreciated...
Orna

[ornam]

re: cufflinks transcript on wrong strand

Just an update - I've since checked four other genes, and they all show the same phenomenon. So, this seems to be a systematic problem.

Orna

[ornam]

cufflinks transcripts on wrong strand - retraction

This is really embaressing - upon examining the cufflinks command I ran (as posted in my original post) I found that I actually ran with a different libraryType than I had intended (fr-firststrand instead of fr-unstranded). When I do run with libraryType set at fr-unstranded the transcripts are on the right strand. So, I am retracting my original complaint.

with humble apologies
Orna