I'm trying to use RNA-Seq data to update gene annotations for Arabidopsis. To start, I tried running CuffLinks over some RNA-Seq alignments.

However, I'm running into a problem. It looks like CuffLinks is predicting ridiculously long UTRs for genes that overlap on opposite strands. The RNA-Seq data are not strand specific.

Please see this link to an igb snapshot that show an example - https://www.dropbox.com/s/10rtzii2w0usjnv/CuffLinksOverlapping.png

Is there a parameter setting I can change in CuffLinks to stop this behavior?

I looked at the user's guide but couldn't find something that looked relevant to this question.