We watched this talk live today (see the YouTube link, below) and highly recommend it. Anne-Ruxandra Carvunis is an associate professor in Computational and Systems Biology at the University of Pittsburgh School of Medicine:

A couple of takeaways:

• Carvunis stressed that the genetic patterns she is studying do not fit with the evolutionary theory she learned “since I was an undergraduate.” Her reaction is mirrored by the de novo gene literature itself, so she’s hardly alone in this response.
• Intergenic regions in yeast look remarkably like “genes-in-waiting,” not random nucleotides.

The audience, as you should be able to tell from the Q&A, was really intrigued, and many were downright surprised. Carvunis of course presented her data and arguments within an evolutionary framework, but there was enough in the talk challenging conventional theory that no one who follows the subject closely should come away disappointed.

The talk itself runs about 30 minutes, and the remainder is Q&A. There are many design implications in the yeast de novo data, which you are invited to consider.