One of the things that has made annotating genes in the maize genome so difficult (there are currently two sets of gene models one with only 32,000 genes, which is low estimate, and the other with 100,000 is far too many) is the presence of large numbers of gene fragments that have been captured and duplicated by a class of transposon called helitrons (yes I know that sounds like a character from Transformers).
The helitron captured fragments are copied from real genes (often multiple pieces are captured from different genes) which is why many gene annotation programs (trained to recongize the difference between genes and non-coding DNA) will identify the fragments being genes themselves.
What if some of those fragments actually are genes? By combining pieces from completely different genes, helitrons could be a whole new source of crazy new genes that natural selection could act upon.
That is the question the authors of this poster are trying to get at, by identifying more helitron fragments and checking to see if those fragments were actually expressed in the genome.
Allison Barbaglia et al. “Accessing the transcriptional activity of Helitron-captured genes of maize” Poster #243 2010 Maize Meeting