Ultraconservation and ultradivergence of human DNA sequences

CREATED: 200805140603 SPEAKER: Shyam Prabhakar, Genome Institute of Singapore TITLE: Ultraconservation and ultradivergence: insights into human development and evolution through statistical analysis of DNA sequences

Use evolutionary conversation to find enhancers, silencers, MARS, etc. non coding regions are harder to annotate

“laws” of sequence evoluation

neutral rate -> nonfunctional

slower than neutral -> shared function

faster than neutral -> adaptive functional change

** ultraconserved elements with zero divergence in human/mouse/rat

Gumby: a general algorithm for quantifying constraint http://pga.jgi-psf.org/gumby/ estimate P-value using permutation test, Karlin-Altschul statistic

Are ultras a distinct class? No.

Distribution of p-values

** similar trend as for all conserved regions

Function of neighbouring genes

** similar as for all conserved regions

Function of ultras

** 50% are enhancers (same as for non ultras)

** ultradivergence: what makes us human

• coding region changes
• cis-regulatory changes

which are the logic elements

which have human specific changes

Which CNSs show an excess of human specific substitutions?

• human specific subst probability at each position
• calculate surprisal of entire CNS
• calculate null distribution of surprisal
• surprisal-test P-value quantifies human-specific acceleration
• found 992 human accelerated CNSs, only 1/734 GO biological process showed significant association, cell adhesion (Prabhakar, Noonan et al., Science 2006)
• rapid changes in regulation of neuronal cell adhesion genes
• top human accelerated CNS was found to be expressed where the thumb joins the wrist