The second day of Biology of Genomes has been and gone, and a brief wine-and-cheese interlude followed by talks to 11pm has ensured that everyone came away with a warm, if somewhat fuzzy, overview of the final session on Cancer and Functional Genomics.
Ewan Birney presented a deluge of information on ENCODE data with a host of interesting insights that I was entirely too jetlagged to take in at all, so will not attempt to discuss. However, there were a few interesting talks that I think I followed well enough to communicate.
As usual, you can follow live tweets from the conference at #bg2011. While poor WiFi performance has lead to somewhat patchy coverage of talks, most presentations have at least one person reporting on them.
Digging deeper into GWAS hits
John Stamatoyannopoulos reported results on high-throughput screening for non-coding functional elements, as part of the ENCODE project. ENCODE has now screened for DNAse1 hypersensitivity sites (DHS) in over a hundred different cell types, annotating the shared and cell-specific regulatory regions of the genome. This assay recovers about 95% of classical enhancers, silencers and insulators. A total of 11% of the genome is covered in such annotations, with about 20% being specific to one cell type.
The really remarkable part of this research is how much light it can throw genome-wide association signals. John presented work showing that 53% of hits in the NHGRI database lie in DHS, and this rises to 63% if restricted to replicated hits. You can see which cells these DHS were found in, and the answers make physiological sense: for instance Crohn’s disease variants are concentrated in T-cells, and ADHD hits in fetal brain cells.
You can dig even deeper, by screening for differential hypersensitivity between the different GWAS alleles to see if the hit is disrupting the functional entity. In 44% of cases that could be tested, evidence of allele-specific sensitivity was found. Many of these seem to disrupt transcription factor binding sites, and by looking at what transcription factors are effected, you can reconstruct regulation pathways that are disrupted. There is a remarkable amount of information here, and it is going to be very exciting to see how much more novel biology can be extracted from the combination of ENCODE annotation and GWAS hits.
Non-coding behavior in coding DNA
An enticing flipside of the growing understanding of non-coding regulation is that coding variation could also be regulating gene expression in a non-coding fashion.
More than 60% of conserved non-coding DNA of the zebrafish act as enhancers. Given this, and previous research showing that exons can drive reporter gene expression, Deborah Ritter decided to see whether conserved coding exons might also do the same.
She made constructs containing conserved exons from 31 developmental genes, combined with GFP and a generic promoter. These were added to zebrafish, which were then checked for GFP expression during development. A pretty sophisticated measuring protocol was used, which involved using a control plasmid lacking the exon to account for tissue-specific background GFP expression. Also, it involved a voice activated iPhone app.
14 of the constructs resulted in strong, region-specific expression during development. In 80% of the assays, the expression of exon-enhanced GTP followed a similar pattern to that of the developmental gene that the exon came from. This suggests that the exons themselves may be in part responsible for the region-specific expression of their host genes.
This is a controversial idea, and more evidence will be required to support it. There are some difficult questions about what the proper comparison test: is 14/31 is the number of random sequences of DNA that would act as enhancers somewhere in development? Do non-developmental or non-conserved exons show this pattern? Can this be validated using an independent technology? However, if true it will certainly change the way we think about exonic variation and gene expression.
Thank you to John Stamatoyannopoulos, Deborah Ritter and Jeffrey Chuang for giving me permission to talk about their work, and in particular to Deborah who gave me a lot of useful comments. The gene regulation image at the top cropped up in a number of talks. I took it from this Kenyon College website