Transcript Populus trichocarpa

Recombination rate variation, hitchhiking, and
central-peripheral structure shape deleterious
load in black cottonwood
Jason Holliday
Department of Forest Resources and Environmental Conservation
Virginia Tech
Understanding the adaptive past to predict the
adaptive future
• The rate of adaptation to anthropogenic climate
change depends on:
–
–
–
–
Number of loci involved
Distribution of effect sizes
Distribution of adaptive alleles in the genome
New mutation vs standing variation
• Many non-adaptive processes also play a role:
– Historical demography (e.g., bottlenecks)
– Migration rate relative to selection
– Deleterious alleles
Populus trichocarpa as model to understand
adaptive potential and constraint
•
Compact (~480mb), sequenced
genome
•
Clonal propagation
•
Wide latitudinal and altitudinal
range = strong differentiation for
adaptive traits
•
High gene flow via wind polination =
low background population structure
•
Postglacial history = possbile
adaptive constraint
Sampling and phenotyping
• 449 samples spanning the
latitudinal and altitudinal
range of the species
– Outgroups: P. tremula, P.
tremuloides, P. deltoides
• Re-sequence most of the
‘gene space’
– Sequence capture
• Replicated gardens in
Virginia and British Columbia
– Phenotyped for growth, bud
phenology, cold hardiness
Genotyping by sequence capture
Sequence capture recovers regions of interest
in a mostly repeatable way
Coverage is tightly linked to bait locations
Genomic sampling
• ~ 2 million single nucleotide polymorphisms passed
quality filters
Adaptation is not just about having the ‘right’
alleles; it’s also about not having the ‘wrong’ alleles
• Many tree species
suffer substantial
inbreeding depression,
while hybrid vigor is
common
• Likely due in part to
accumulation of
deleterious alleles
Wang et al 2004
Deleterious alleles
• Coding SNPs often simply categorized as synonymous
(same amino acid) or non-synonymous (different amino
acid)
• But not all non-synonymous SNPs are equivalent
≠
Ala
Gly
Ala
Cys
Factors that may govern accumulation of
deleterious allele
• Genomic
– Recombination
– Hitchhiking
– Direct positive selection – are some of these alleles
conditionally advantageous?
• Demographic
– Population history (bottlenecks, etc)
– Population size (efficiency of selection)
Sorting Intolerant From Tolerant (SIFT)
• Uses multiple alignments from related species to
classify SNPs as tolerated or damaging
• Examples:
– If no amino acid substitution is found at that site, any nonsyn change is deleterious
– If only hydrophobic residues found, changes to other amino
acid classes deleterious
• No information about protein structure, but performs
similarly to algorithms that account for this
– Advantage: can be used on just about any sequence, not just
those with known structure information
Deleterious SNPs segregate at lower frequency
than tolerated SNPs
Derived allele frequency
Deleterious : tolerated ratio
Percent
Deleterious SNPs segregate at lower frequency
than tolerated SNPs
Derived allele frequency
Determinants of deleterious frequency:
recombination
More generally – recombination rate variation correlated with deleterious ratio
Determinants of deleterious frequency:
hitchhiking
• iHS = measure of
incomplete hitchhiking
events
• iHS higher for
windows enriched for
deleterious SNPs
Density
• Deleterious ratio
higher in top 1% of
iHS windows
60
40
All
Deleterious enriched
20
0
• Signal of direct or
linked selection?
0.6
0.8
1.0
Normalized iHS
1.2
Any evidence for direct selection?
• Suggests deleterious
alleles mostly not direct
targets of positive
selection
0.15
0.00
• Deleterious alleles
underrepresented among
FST outliers
0.05
0.10
• FST outliers identified
across three sampling
transects (two
altitudinal and one
latitudinal)
All SNPs
Significant outliers
Coquihalla
Highway 99
Deleterious:Tolerated ratio
Rangewide
Determinants of deleterious frequency:
historical demography
• Like many northern
species, poplar
experienced bottlenecks
associated with
Pleistocene glaciation
– Increased drift in leading
edge populations
• More generally, variation
in Ne across the range
may lead to differential
efficiency of purifying
selection
Determinants of deleterious frequency:
historical demography
Why does this matter?
Deleterious alleles affect fitness
BC Garden
VA Garden
Can we detect the signature of deleterious alleles at
individual SNP loci?
• Sort of – deleterious alleles overrepresented among associated genes,
but underrepresented among associated SNPs
• Low frequency deleterious alleles generating ‘synthetic’ associations?
Conclusions
• Deleterious alleles impact performance, and may explain some
phenotypic associations for tolerated SNPs
• Accumulate as a result of recombination rate variation, hitchhiking,
and demographic history
• More common in peripheral populations
– Response to selection may be weaker in these populations, which suggests less
adaptive potential under climate change
Acknowledgements
•
•
•
•
Mandy Zhang, Lecong Zhou, Haktan Suren, Rajesh Bawa
Kyle Peer, Clay Sawyers, Debbie Byrd (VA Garden)
Mike Carlson (BC Forest Service) and Cees Van Oosten (BC Garden)
Carl Douglas and BC Forest Service – access to some BC samples