Transcript Ultras (cont.), hCONDELs (Gill)

Any Project reflections?
http://cs273a.stanford.edu
[Bejerano Fall10/11]
1
Lecture 16
Ultraconservation cont’d
hCONDELs
http://cs273a.stanford.edu
[Bejerano Fall10/11]
2
Ultraconservation as a Phenomenon
We do not see a bump in the curve
Few species
More and more species
Hmmm….
http://cs273a.stanford.edu
[Bejerano Fall10/11]
3
Ultraconserved Elements: What we don’t
• What maintains so much conservation?
*
*
http://cs273a.stanford.edu
[Bejerano Fall10/11]
*
*
*
4
Broad Guess
It’s about 3-D structure.
Observation: rDNA (18S, 28S) have ultraconserved stretches,
multiple constraints in a complex 3-D structure, the Ribosome.
•ncRNA ultras: structure confers function
•Splicing related ultras: the Splicosome
•Cis-reg ultras: TSS 3-D proximity, chromatin
and/or packed TFBS (Transcription factories?)
TSS
http://cs273a.stanford.edu
[Bejerano Fall10/11]
5
How special are the Ultras?
election
Ultraconservation
http://cs273a.stanford.edu
[Bejerano Fall10/11]
6
Question
Is functional density (conservation %id)
a good measure of dispensability (deletions)?
“Average” conserved element (85%id to mouse)
Ultraconserved element (100%id to mouse, rat)
Both of these regions are clearly under purifying selection.
Which one is more likely to be lost throughout evolution?
http://cs273a.stanford.edu
& Bejerano, Genome Res., 2008]
[Bejerano [McLean
Fall10/11]
7
Most all Non coding sequence under purifying
selection clearly resists deletions.
•Loss rate nearly constant regardless of conservation %id
• Similar loss rates for slowly-evolving non-exonic sequence and
coding exons suggest functionality of non-exonic sequence
Rodent-Dog Regions Lost in Primates
Slope of best-fit line = 0.00006
http://cs273a.stanford.edu
[Bejerano Fall10/11]
8
If functional density doesn’t correlate
with dispensability, what does?
Deep orthology is a better predictor of persistence
Rodent losses
http://cs273a.stanford.edu
Primate losses
[Bejerano Fall10/11]
9
Genomic Distribution of Ultraconserved Elements
•exonic
•non
•possibly
http://cs273a.stanford.edu
[Bejerano Fall10/11]
10
Outliers: Ultraconserved Element uc.338
uc.338 paralog family
http://cs273a.stanford.edu
[Bejerano Fall10/11]
11
Coelacanth Homologs Closer than Human Ones
[Bejerano et al., Nature, 2006]
http://cs273a.stanford.edu
[Bejerano Fall10/11]
12
Uniquely Abundant in Coelacanth
Upto 80%id between Coelacanth instances
and some human instances, inc uc.338.
?
x
100 diverged copies in a Gigabase
60 highly similar copies in a Megabase
http://cs273a.stanford.edu
[Bejerano Fall10/11]
13
Cis-reg & Ultra elements from
Co-option event,
probably due to
favorable genomic
context
obile Elements
All other copies
are destined to
decay over time
at a neutral rate
[Yass is a small town in
New South Wales, Australia.]
http://cs273a.stanford.edu
[Bejerano Fall10/11]
[Bejerano et al., Nature 2006]
14
The LF SINE (for Lobefin Fish / “Living Fossil”)
not similar to any known repeat
out
Reconstruction
back
target site
duplications
http://cs273a.stanford.edu
[Bejerano Fall10/11] [Bejerano et al., Nature 2006]
15
Exapted Into Which Cellular Roles?
No evidence for Transcription (Tx) as small RNAs,
no orientation preference in introns, not in antisense Tx.
?
Human instances cluster together,
found <1Mb from 35 TFs (P<3*10-6).
x
http://cs273a.stanford.edu
[Bejerano Fall10/11]
16
Repeat made Regulatory Region
in situ
Conserved
Element
Minimal Promoter
Reporter Gene
transgenic
http://cs273a.stanford.edu
[Bejerano Fall10/11]
17
UC.338 comes from an ancient repeat
ultraconserved
exon
novel
coelacanth
repeat
enhancer
LF-SINE
[Bejerano et al, Nature ,2006]
http://cs273a.stanford.edu
[Bejerano Fall10/11]
18
A lot of conserved elements come from repeats
5%
[%age
of H.G]
repetitive
conserved
1.5%
20%
>100Mya
highly conserved non-coding
(think functional, regulatory)
50%
>100Mya
mobile element
instances
10,402 elements!
1Mb, avg len 100bp
(But no other ultraconserved element to date)
http://cs273a.stanford.edu
[Lowe, Bejerano &Haussler, PNAS ,2007]
[Bejerano Fall10/11]
19
Ultraconserved Elements: What do we know?
•
•
•
•
•
•
•
•
Excessive sequence conservation exists.
Set is heterogeneous from a functional perspective.
Four can be KO-ed with no clear phenotype.
Yet, the set is under extreme selection in natural
populations, both for mutations and deletions.
Most ultras have deep orthology, and no paralogy.
One ultra comes from a mobile element co-option events.
Others may have come from similar events.
Ultras appear the tip of a continuum, not a unique peak.
http://cs273a.stanford.edu
[Bejerano Fall10/11]
20
hCONDELs (pre-pub., slides omitted)
http://cs273a.stanford.edu
[Bejerano Fall10/11]
21