Genetics of Larval Fitness in the Pacific Oyster
Articles Blog

Genetics of Larval Fitness in the Pacific Oyster

welcome thank you so much for
tuning in we’re pleased to have you join us for the California Current
Acidification Network Ocean Acidification Roundtable Discussion for
July 2019. The title of today’s presentation is genetics of larval
fitness in the Pacific oyster responses to acidified seawater and temporally
dynamic selection processes. This series is hosted by the California Current
acidification Network in short the series is intended to create a dialogue
among industry members natural resource managers and scientists within the
California Current ecosystem with awareness of and access to the research
and its applications and uses the participants will be able to collaborate
to better understand and adapt to ocean acidification moving forward. Evan, if you could advance the slide. Oh
that’s right I’m Teri King with the Washington Sea
Grant program and chair of the California Current Acidification Network
steering committee and I’ll be moderating today’s session as well as
running the logistics behind the scene. During the presentation attendees will
be in listen-only mode. You’re welcome to type questions related to technical
issues or questions for the presenter in the questions box at the bottom of the
control panel on the right of your screen. I will be monitoring incoming
questions and will respond to them or pose them to our speaker after their
presentation. We’re also recording this session and we’ll share the recording on
the California Current acidification that website in the future. We’re very
excited to have Dr. Evan Durland speaking with us today all the way from
Sweden. We have a number of other international researchers on the call
today I believe our furthest is the Galapagos which will be very interesting
to see if her connection holds through the entire presentation. Dr. Evan Durland
is a postdoc at the University of Gothenburg and Sweden and a former
student of Chris Langdon’s. His current project is focusing on optimizing the
farming of mussels and kelp as bioremediators with semi-enclosed salmon farms in
Norway. Evan received his Bachelors of Science from Colorado State University a
master’s in aquaculture from Auburn University in Alabama and his PhD in
fisheries from Oregon State University just a few months ago. In
between degrees he also worked breeding oysters for the molluscum brood stock
program and for private aquaculture companies in Indonesia growing both
grouper and pearl oysters. Evan, I’ll leave everything to you, you have
complete control. Evan, could you speak a little closer to the microphone, project a bit? We’re having a hard time
hearing you could you speak louder or closer to the microphone so some of these physical characteristics make it very attractive to aquaculture. Most widely farmed species today. This resilience to
environmental stress and these characteristics have led this species on
something of a worldwide tour, it’s native to the western Pacific there we brought it over
in the 20s and subsequently this species has been transplanted to every continent on the globe. The places in green
here are the places where it’s colonized. where the point is that the species have been successful in a wide range here in the West Coast we found the
pacifically strip across this range the most substantial commercial it’s a pretty significant culture
industry on the west coast accounting here and really shellfish community – yep Evan could you maybe put in the we
didn’t I will apologize to everyone we tried the mic and it ended up working
fine and now it seems like it’s not can you plug in the earphones again and see
if that helps you got it okay we think that part of this is though you know
people in Sweden are all online obviously it’s 9 o’clock at night and
maybe it’s an issue of the wire and the way that we’re connecting can you hear
me any better now yes okay so we’ll go with that
all right so we’re all pretty familiar with the adult oyster but what I’ll be
talking about most today is the larval phases of this animal which to me are
the most interesting part of its lifecycle. So it’s a successful broadcast
spawner as we all know so it sits in one place for all of its life and emits
sperm and eggs into the water column where they hopefully meet and fertilize
to produce an embryo and then it goes through a series of developmental stages
during its larval fades and I’ve highlighted some of the stages here this
first one on the left here D-larvae is the first shelled veliger stage so
shortly after the fertilization within 24 hours the animal undergoes a
dramatic transformation to become this fully swimming and feeding veliger larvae.
This is also the phase of initial calcification so they create their shell
very rapidly at about 12 to 18 hours post fertilization. They’re this full
organism by 24 hours. Then for the next two to three weeks they continue to grow and feed and acquire nutrients through these
veliger stages up here until they reach this final swimming stage called the
pediveliger it’s called pediveliger because they
actually grow a foot that extends out of their shell. They start searching for
substrate someplace they want to settle down for the rest of their life. At this
stage they undergo a final transformation from swimming larvae to this juvenile
spat so this again is a very dramatic transformation and they rearrange a lot
of their body and their organs and become this final juvenile oyster
that we are more familiar with so one of the things that has really
been a key component of increasing the success of the oyster across the world
has been a hatchery production seed so initially all stocks were supplied by
naturally recruited populations both in Japan as well as here in the west coast
and Washington but after a series of years where that recruitment was a bit
spotty we developed hatchery techniques to grow these things in optimal conditions
and expand their their adult growing region substantially so the second thing
that growing things in a hatchery enclosing the life cycle allows is for
an improvement of the organism so starting in 1995 Chris Langdon started
the molluscan brood stock program here at Oregon State University and for the
past 20 years we’ve been breeding oysters for improved field traits at
farm environment so that’s growth and survival in estuarian environments across
the Pacific coast. Currently in the seventh generation of selection and have
substantial gains over the wild stocks and a lot of environments
but the main challenge for oyster aquaculture in the past decade has been the aspect
of ocean acidification and I think again most people sitting in tonight
today are familiar with OA just briefly it’s the process whereby we’re
dumping CO2 into the atmosphere it’s getting absorbed by the oceans and
making it more acidic which makes it do more difficult for marine calcifying
organisms to create shell and survive in these environments. On the right you see
the IPCC projected atmospheric CO2 levels and the takeaway here is it’s a
problem for us now where we’re sitting here around this 400 mark and this
problem is only going to get worse through the end of the century so in the
best case scenario if we stop emitting carbon dioxide we’ll still see it climb
up to just under 600 parts per million in the worst-case scenario this could
climb up to a thousand parts per million so it’s a problem that we need to be
getting our heads around and getting a head out now
so it’s a global problem but we’re feeling those effects more acutely here
on the west coast due to natural weather patterns and so we get these summer
North wind’s that drive coastal waters offshore they’re replaced by deep ocean
water masses that are rich in carbon so we get pCO2 levels that climb up
above 1,100, 1,200 and we get pH levels that drop below 7.7, 7.6
every summer now This has been a significant problem for the shellfish
aquaculture community on the west coast and since 2007 it’s been accountable for
significant losses in the hatcheries so some years up to 75% reduce production
and losses up to 10 million dollars a year for these small businesses. Overall
the production of gigas has decreased in the past decade although we’re seeing
some rebound and so it’s a big deal for the West Coast and how to adapt
to these new environmental conditions Now what remains a little bit less well
known is how these conditions affect natural stocks so for example Willapa Bay.
We know that they’re exposed to upwelling in summer months and they
experience acidified sea water conditions but it’s difficult to tie the
naturally stochastic patterns of spawning and recruitment with these also
stochastic patterns of upwelling but we have good reason to believe that these
are not helping those wild stocks as well and they’re being exposed to these
acidified conditions as well as the aquaculture populations.
All right so to summarize 10 years and an extensive amount of research
in one slide so this has been something of interest for a lot of research groups
to understand how these conditions specifically are affecting the biology
of shellfish, specifically oysters, and one of the big takeaways it has to do
with calcification so like I said at this early development stage they’re
forming that first shell very quickly and when they’re fully exposed to the
seawater conditions so this is some iconic work from George Wald Bussard and
Alan Barton demonstrating that the association of early
larval development and aragonite saturation it strongly dictates the
success of these larvae so you can see on the top here on the left is percent
normal that’s the percentage of these d– larvae that have a normal shape
which is depicted on the bottom in the picture and on the right you have larval
size and in both cases you can see as we decrease our aragonite saturation state
we get fewer normal and those larvae are smaller. So this is dramatically
impacting the ability of these animals to calcify in
these conditions. So a lot of research effort has been focusing on this early
development period and how calcification and larval processes are affected in
acidified conditions but the rest of the life cycle there has received relatively
less attention and that’s for good reasons partially because it’s difficult
to grow larvae all the way through to settlement also because that settlement
is very difficult to capture and quantify. So we have a lot less
information on the long-term or chronic effects of acidification on larval
fitness and furthermore we have extremely little knowledge on what
genetic consequences these stresses have so we anticipate that this is a pretty
serious stress on these organisms but we haven’t yet really grasped how that
might act as a selection pressure to change the genetic composition of some
of these larval groups for this organism more applied we don’t know if there’s
differential sensitivity between stocks in the northwest. There’s some evidence
from Australia that different stocks are differentially sensitive to OA. On the northwest we don’t have any kind of baseline reference whether
stuff in Willapa or Puget Sound differs in its sensitivity to OA relative to
something like the molluscum rootstock program. So all that sets the scene for
the three main subjects I’d like to talk about today so the first part will cover
some of the research I’ve been doing with looking at the phenotypic effects
of ocean acidification and larval oysters so that’s the chronic long-term
effects on development and recruitment to spat. Specifically also is there a
difference between larvae that come from those selected lines or MVP and those
that come from naturalized populations in Willapa Bay.
The second part we’ll take samples from the same experiments and look at the
changes in genetic composition of those larval groups in ambient and certified
seawater. So our questions are does larval culture and acidified
conditions have genetic consequences on those animals and how are they
similar or different between MVP and Willapa and then for this last section
it’s really diving more into the dynamic temporal changes in genotype frequency
in these groups during larval development. There’s been some work
looking at the effect of genetic load on these populations and we have now data
that can kind of resolve that with better resolution looking at how some of
these changes take place during development.
Alright so jumping straight in so the experiments that I conducted which are
the framework for all this data were conducted in 2015 and 2016 and in the
beginning our goal was simply to compare MVP to Willapa so to that end we took
Willapa wilds from the Missal region in southern bay in 2015 and brought those
into the lab for a year. We brought them in in 2014 for a 2015 spawn, brought
those into the lab, conditioned them alongside MVP stocks so we pulled our
top families from the fifth and sixth generation and we took a lot of those
families to create a diverse pool we’ll talk about a second. we repeated this
experiment in 2016 and got oysters from North Bay which is more exposed to
upwelling up here in Stoney Point region to replicate the experiment. So in both cases
we brought all these broodstock in and our goal was to create a big diverse
gene pool for experimentation and this is important because we know that
there’s likely to be some differences between individual families and we
really wanted to homogenize that effect so we created 95 crosses from each of
these stocks in each of these years to get this kind of mean fitness for the
entire population and in the case of MVP because we have a pedigree we can
actually recreate that pool so for 2016 spawn we went back to the same families
and remade that pool as close as we could to really kind of replicate the
genetic stock and those experimental conditions.
So for this we use static culture. We use these ten liter polycarbonate buckets
with a screw lid and a rubber seal and we used pCO2 levels with ambient at 400
parts per million and high pCO2 at 1600 now if you’re familiar with the
literature 1600 is pretty high for OA stress but it’s also realizable in the
west coast every year we see acidification levels
getting up to 1600 parts per million so we consider it a relatively high but
realistic stress for these animals we reared them for about three weeks 22
days in 2015 and 24 days in 2016 and changing water every other day to keep
the conditions fresh. In 2016 we added a sixth replicant to each of these levels
to to fortify our statistical string and like I said our goal here was not
only to look at that initial development but to look at the long-term effects so
we have samples throughout the larval development period including through settlement in those spat populations and we’re looking at total number surviving,
the size of those individuals, their developmental progression and of course
also their genetic composition okay so jumping into some of the results
this is kind of the high altitude view of the results from both these
experiments so we have 2015 on top and 2016 on bottom and this is the total
larval survival and the first thing you will notice on this graph is that there
doesn’t seem to be much difference and that’s kind of the point that when you
look at the larval phenotypes from a high altitude and say what’s the overall
outcome you can’t detect much of a difference and it really involves
breaking it down into these individual phases of that larval development to
understand how these conditions are affecting the larval physiology and what
the differences are between these two broodstock groups. So we’ll be talking
about this initial phase or early development also the middle phase or
veliger stages and then finally this metamorphic transition to spat
so starting with that early development period so from day 0 to day 2 or 48
hours post fertilization these are some of the typical metrics we see looking at
OA work. On the Left we have survival, on the right we have percent normal and the
takeaway here is that we didn’t have an overall effect of acidified conditions
on a total survival in fact it was gently positive in each case we had
slightly more larvae in those acidified conditions but what we did see was that
the normality was significantly reduced this is the somewhat classical response
to OA conditions where we get this lower percentage normal and in each year we
saw for reduction those percent normal so to kind of summarize these effects
high pCO2 had a relatively consistent increase in total survival but a
decrease in the percent normal. The broodstock effect here was variable and
really not that significant so it was kind of up in one year for MVP and down
in another and no real consistent effect on normality so we don’t have a strong
broodstock effect for these early phenotypes as we move into these veliger stages
the striking thing is that all the differences we saw early on disappeared
in both years we had very consistent survival and growth in our conditions
through to that pedo-veliger stage so we saw the same amount of mortality and
growth both in MVP and Willapa low CO2 high CO2 so no real changes that were
apparent earlier on carried through to this veliger period. As we approach
this pedo-veliger period we go through that last transition to spat and this is
where things start getting interesting again so here is total
larval survival that’s the survival of every of the larvae on the left and the
settlement success on the right and that’s that those that were from went
from pedo-veliger to spat so what we saw here was striking differences
between both these experiments. In 2015 this survival was not affected by … culture and there wasn’t a real big difference between MVP and
Willapa. The total spats the success also wasn’t that different between low CO2
and high CO2 but MVP did appear to have more spat. In 2016 we saw what would be
maybe the more expected response and that’s that we had a much lower survival
of larvae down here and the settlement reduced, about 42% less spat coming out of
those high pCO2 cultures. So again to summarize we have MVP having no real
effect on survival across both those experiments but a consistently improved
settlement success high pCO2 culture had this weirdly variable
response to this performance metric where in one year it didn’t affect them at
all, the next year it had a dramatic negative impacts. So the struggle at this
point was to understand how those two things fit together and the real clue
here was looking at the nature of the mortality and this is again an important
component is understanding how things are dying in your culture not just how
many are dying and so in this case and when we look at the samples from 2016 at
the final time point we see that all the mortality that we saw 99% was
happening in these underdeveloped larvae which is depicted by these kind of white
shells here and not in the pedo-veligers and not in the spat so that’s an
important distinction. What that means is that the larvae weren’t trying and dying
it’s that they were withheld from getting to that settlement competent
state where they could even try and settle out and that fits with some other
work from colleagues down in California looking at the metabolic costs of
survival in high CO2 environments so this is some work from Freder last year
and looking at the metabolic cost and the metabolic partitioning in these
conditions. So the top graph really demonstrates the efficiency of these
animals in high CO2 environments. This is the protein depositional efficiency as you
can see as you get to high pCO2 which in this case was around 1,100 they get
much less efficient at creating new protein and this bottom pie chart
demonstrates two different strategies so in control CO2 you have 66 percent that
goes to protein synthesis of datp and the remaining is for maintenance as
you get to middle CO2 they pump up the protein but if you get too high pCO2
they increase the total metabolic pool so they’re requiring more energy and
their efficiency goes way down. So those observations fit with what we saw in
2016 in that these animals simply just couldn’t keep up in this high demand
environment. And the other corroborating bit of evidence comes from the brood
stock so one thing we saw straight away in 2016 was that the condition of the
brood stock was much worse. On the top here you can see the total number of
eggs we got out of each female on average
up here we had 300,000 per female and then down here we had considerably less. We can just see that from the broodstock they were poor condition. Now curiously that didn’t equate into energetic content of the eggs we did
lipid analyses and actually the eggs in 2016 had more lipid than they did in
2015 so it’s not simple energetics but something in there about the egg quality. You’ll notice we don’t have Willapa for 2016
that’s because the samples got contaminated but what this experiment
taught us was the Oh a is certainly an important factor in dictating the
performance of larvae and it’s an important factor for the chronic effects
but it’s not the only one and it interacts with numerous other stressors
right so it’s interacting with energetics as well as temperature which
other papers have shown us also things that are happening in the natural
environment like hypoxia and diseases and micro flora blooms. So really OA is one of the major stressors but we have to integrate that
into a broader context of all the other things that are happening to the larvae
at this time No as you can imagine there’s a lot more of
the discussion part of this project and you can see that in our new publication
it maps if you care to look it up but this was an interesting takeaway from a
messy dataset where sometimes when you find things you didn’t expect to see in
can paint something of a richer picture So to summarize the phenotypic effects
of growth in these OA environments what we see is that high pCO2 here had
very consistent effects in early development so we had a general
improvement in survival and a significant decrease in percent normal
but we had some variable effects of high pco2
over the seventh period however despite all that MBP continued to produce more
and bigger spat in both those conditions in both those years MBP had 55 percent
more spat in ambient and 37% on average more spat happy co2 environments they
were also bigger so 5% bigger in ambient and 23% in die co2 so that was an
interesting takeaway that MBP had maintained a performance advantage
across that amount of variability so that brings us to this next question if
this has to do with genetics from some unintended source with our rootstock
program what can we observe with changes in
genetic composition of these larvae in both normal and acidified
environments so to get at that question we took
samples from our 2015 trial and we’re looking at the overall change from start
to finish in these groups so at day two and day 22 and we’re looking at this
total genetic change over development in both those conditions we can
conceptualize this in another way and that’s that we have this big diverse
gene pool going into this experiment so 95 crosses in each of those groups and
the question we’re kind of asking is how do those compositions look in ambient
and how they look in high co2 environments and are those changes
consistent between our selected lines and our wild stocks
so traditionally the way you devalue ate this is by genotyping a good number of
individuals and when you in genotype and individual you get their genotype so you
get two alleles in this case this is a heterozygote and as a or B or a and B or
homozygote as a a well in the case of oysters they’re quite small and
difficult to sequence individuals and as as of yet you can’t really get enough
DNA out of them for modern sequencing platforms so we adopted a pool seek
approach and this is where you take a bunch of larvae and you extract all
their DNA and your sequence all of that DNA so the drawback from this approach
is that you lose individual genotypes so you get this this whole mix of different
alleles and you can’t call genotypes you call minor allele frequencies in this
example we would observe 6a alleles for B alleles and we would calculate a minor
allele frequency or the frequency of B at 40% so for the rest of this we’re
gonna be talking about minor allele frequencies and that’s what that means
is that among that entire population how many of those alleles do we see
so we had 1288 snips across the genome that met our coverage thresholds I won’t
go too much into detail and the methods if you’re curious I’m happy to talk
about all my trials and tribulations figuring about how to use the pool
secret but essentially we took all these markers these individual nucleotide
variation and we modelled it in terms of the developmental stage so day 2 or 22
the treatment which is low or high co2 and the interaction between those and so
this is one way we look at that data and so on the top we
have just ambient conditions over the two different time frames so we have
just day two samples here on the x-axis and just day 22 samples in the y-axis so
what this graph is showing you the scatter plot is the starting frequency
on the x-axis if it changed nothing at all
it would lie on this 45 degree line any distance from that line is the
distortion in allele frequency either up or down by day 22 the highlighted dots
are those that were found significantly different by that model after false
discovery correction so these blue dots are those that are changing under
developmental processes and on the bottom we have a different comparison
that’s the comparison of just spat but between ambient and high pco2 conditions
so on the Left we have MVP and the right we’re Willapa and what we see here is
that we get substantially more distortion from those wilds in those
high co2 environment now these Purple’s are those interactive snips look at that
next slide so when we look at all those changes to
altogether what some of the first takeaways are a that Willapa is changing
more than MBP so they had 26% more markers that were changing in ambient or
control conditions than MBP but they had 223 percent more they were changing in
high co2 conditions so there they were much more genetically affected by those
high co2 environments than our selected lines and that alone is a pretty
interesting takeaway but there’s another aspect that was really eyebrow-raising
for us and if you look at the right here this is a cross tabulation of how each
of those markers fits out with the other group so we have will up on top and MBP
on side here for example Willapa for stage effects are just developmental
effects we had 145 markers that change significantly for that group if you look
up here in the red box only four of those also changed in MVP for that same
reason and that’s true for each of these categories as you go down so we have
development high co2 exposure the additive that’s both them together and
the interactive here and across that diagonal you see very very few markers
actually were consistent between these two groups so this was a very big
surprise as well we would fully expect that that the changes in one group
should be reflected somewhat at least in the other group but that was not the
case here now another way to look at this is our similar regions of the
genome changing so this could be that yes yes they’re not the same markers
changing the same way but it all kind of translates back to some maybe arm of
chromosome that’s changing so when we look at this linkage analysis so this is
based off the linkage Maps from Dennis hedgecock we can see these changes are
more or less evenly dispersed across the genome so it doesn’t appear to be the
case that they’re changing in consistent regions of the genome the changes in
will open are much more profound you can see that from their p-values here and
they’re not consistent with what’s changing in MVP so
further way that we can try and make these make sense and be more similar and
that’s that we can look at their functional their putative function of
the genes around each of those markers this is functional enriching analyses
and conceptually what that is is we take a marker which is the star here we go
for a search on the annotated genome upstream and downstream looking for
genes in that region and then we take those genes and put them into a program
say okay are there more of one function than other functions being represented
from those markers that are changing so we used mine J for this and this is the
takeaway from these developmental effects so what you can see here is that
we have very broad processes for developmental effects that’s protein
modification helicase anti on binding metabolic processes and the important
part here is twofold one very few genes fall into those categories so they’re
very poorly represented and they’re also very broadly functional so this is a
score from zero to one of how specific it is with higher numbers being more
broad so for developmental processes what we see is that they’re affecting a
large number of processes and both these groups now contrast that with the same
analysis on those that are changing for high co2 environments what we found is
something very different in this case even though there’s fewer markers that
were affected by those environments they’re much more strongly
over-represented by specific components of physiology in this case it’s
membranes so with integral component of membrane membrane part and just membrane
in general and again these are many more genes that are ascribing these these
categories they’re much more specific and those p-values are stronger so that
was a very compelling takeaway this fits also with what we know from other work
about the importance of membranes in survival and acidified environments so
we have a number of other groups looking at urgent and oysters demonstrating the
transmembrane transport of ions is one way not only do they maintain
homeostasis within the cells but it’s also a way that they foster
calcification so this figure here is modified from paper we
did about shale formation and in with some gene expression work we also
identified transmembrane proteins as one of the upregulated groups in high co2
stress this last paper here ramish they actually showed empirical evidence that
they were doing this so what they did is they took a muscle artery suctioned it
under a pipette and inserted a probe right in that area of calcification
demonstrated that the organism is actually modifying this calcification
and this calcifying fluid so this space where the calcification is taking place
they’re making it more advantageous for the deposition of calcium carbonate so
all these activities have to do with trans membrane transport and that’s been
shown empirically and these other works but what this data kind of adds to that
picture is not only they up regulating or increasing the performance but these
functions are also selective because we’re seeing changes in the genetic
composition of these groups this means that this is a selective pressure for
these activities so what we saw overall was MVP had less
genetic change overall than Willapa and that’s consistent with what we saw from
these survival trends we saw increased survival and settlement success so
collectively that provides us a little bit of evidence for domestication in MVP
lines meaning that they’re surviving better and changing less probably
because they’ve been selected for this kind of uniform stable hatchery
environment the disparate changes are a bit of a mystery they imply either that
there’s different physiological processes being selected in each of
these stocks or that they differentiated enough from one another that the markers
are relevant to one another the encouraging bit is that under those high
pco2 conditions the genetic changes do appear to have some functional
similarity even if the individual markers changing or different and then
the last thing I’d like to highlight on this section of the work is to remember
that we saw in this year’s experiment 2015 no a significant difference in
survival in high co2 environments but nevertheless we demonstrated there
significant genetic change and in Willapa there were two times more
changed so what this means is even if we’re not seeing a mortality then there
are having effects in that population that might carry on to future
generations furthermore it means that there’s probably some trade offs so as
some things are dying in ambient conditions that are surviving in a way
you’re getting a shift in the Fitness Optima okay so that is kind of the overall
effects but another big part of this this puzzle is what are the temporal
changes that are taking place in these larger populations across development so
kind of asking when and how did these genetic changes occur so we recall from
that last bit a full quarter of all the markers we queried showed some signal
change over development and that’s quite a bit when you think about it 25 percent
of all the markers across the team ailments showed a signal change this is
consistent with some of the estimates from previous work by those plowed and
it’s Hedgecock looking at the effect of genetic load on larval survival and
genotype frequencies and larvae and over a series of papers they demonstrated or
suggested that there’s 11 to 19 deleterious low-side present in the
oyster genome that renders around 90% of all larvae genetically and viable now
that’s a staggering statement but it is borne out with the genotype distortions
and what I mean by genotype distortions is kind of depicted here on the right so
for this work they used a lot of microsatellite markers and the genotype
individual spat and compared that to the genotype combinations that would have
come from the parents so on the Left bar you see the genotype composition coming
from parents so in a heterozygote cross you would get 25% of the AAA genotype
25% of the BB genotype and 50% of the a B genotype
at the end of the larval period at day 22 what they observed was a distortion
in that frequency so in this case we reduce our BB genotype and that suggests
that those alleles are detrimental to those animals because they’re causing
disproportionate mortality that genotype now in their 2011 paper they also did
some temporal analyses but probably due to the difficulty in amplifying DNA from
individual larvae the number of markers that were available for these
intermediate larval stages were limited so the work that we did kind of expands
upon that work and offers some high resolution examination of patterns of
change so to kind of put it into framework that’s understandable what the
question we’re kind of trying to ask is yes we know this is the starting and
frequency but can we fill in some of these mystery boxes is what’s happening
in between linear or unpredictable or is it something other
so to get at that we have samples from that same experiment in ambient in this
case we used Willapa samples so while – well genotypes and we have samples that
start at the fertilized egg with day to day 6 8 10 16 and 22 and so we’re
evaluating the change in minor allele frequency at all those time points in
this case because we have a large number of samples we do two kind of data gap
overlap we’re left with 867 snips across the genome and in this case we modeled
it on a simple model which is just looking at the effective age and
importantly we coded age as a factor another linear variable because as
you’ll see a lot of the changes aren’t linear so we had to let that be a factor
so after false discovery correction were left with 516 snips that were
significantly different one or more time points so this is a big messy dataset
and I had to develop some different tools to kind of parse some of the
signal out of here and one of the ways we did really we did this was by using
parametric tests and using what I call sequential pairwise – key tests so
essentially what that does is using this as an example if you have a snip the
change in frequency across time we go back and do pairwise tests to understand
that what was was there a significant change between any two time points and
there’s a couple different scenarios you can get from this test and this first
one is an example of what I call a gradual snip so from start to finish
it’s different but sequential changes are not significantly different so this
has a gradual positive change another option is that you have something that
we call the unidirectional snip so this is something similar to gradual but a
bit more significant where you have one or more time points where that changes
so this was would be during that final metamorphosis you got a significant
change in the frequency the third category is called a bi-directional
change or I like to call them flip-flops and that’s where we get a significant
change that goes in two directions so if we start at 50% through early
development it drops in frequency then abruptly
switches direction and goes in the opposite direction so these are the
flip-flop snips so we have those three scenarios depicted here on the top and
then we also have the tangent to those curves on the bottom so this is the
change in minor allele frequency and so you can see gradual is just kind of
pretty consistently positive unidirectional is consistently negative
with one big jump down in the bidirectional is this more dynamic
pattern of change so we have all these different patterns and we had to use a
different technique in order to kind of parse some of the categorization out of
that for that we used k-means clustering analysis now k-means clustering analyses
is very complicated but conceptually pretty straightforward so this is our
simulated data this is just more of the same simulated data and conceptually
what k-means clustering is going to do is use some math to separate things out
so that they’re more similar in individual groups so in this
hypothetical example we would cluster out these patterns into these three
clusters cluster one is all those bi-directional cluster two is those
gradual and cluster three is those unidirectional snips so this is the
simulated data this is the real data so you can imagine how intimidating that is
to look at the first time but the clustering analyses do a really good job
of teasing out some of the consistent patterns so in this case it worked out
ten specific clusters and this is the change in minor allele frequency for all
five hundred and sixteen of those significant snips and this is the
tangent to those curves so this is the change in minor allele frequency and so
what we can see here is that in these low clusters which are highly populated
we get a little wobbling around that line so you get some significant changes
but they’re not a really dramatic as we work our way up this cluster number what
we see is they’re getting more and more dynamic and swings and more and more
pronounced and so when you start getting here you get these overall patterns that
are swinging in frequency across those developmental periods especially up in
these high order clusters right so when we compare these two methods
they agree pretty nicely so in these low categories with the light wobbles we end
up seeing a lot more of these gradual snips as we go up in cluster number we
see an increased prevalence in these bi-directional snips so they they agree
pretty well so that’s a pretty confusing graph and I
understand that because it took me six years to figure it out so there are some
pretty interesting implications to these flip-flop patterns of change in allele
frequency and I’m gonna try and put it back in frame in the framework of
changing genotypes so this is that same example of before that came from the
plows 2016 paper where we have this change from 50% minor allele frequency
to 35% and in that case they knew the genotypes so they could give us that
change in genotype frequency well in our context we’re looking at minor allele
frequency so there’s a number of ways I can sort of solve that problem with
minor allele frequency that could be either of these linear solutions or this
gradual solution then all three of these make sense that at some point in life
that BB genotype was disadvantageous and dropped out of frequency but these
flip-flop patterns imply that there’s also a good proportion of these these
markers that are changing in patterns like this and so if we translate that to
changes in genotype frequency and we can do that it looks more like something
like this and so the implications here are very different where before we said
well it’s the BB genotype that is negative or deleterious in this case
what we see is that we have a prevalence of that AAA genotype that eventually
gets reversed and then flips back so this pattern of selection leads to very
different conclusions then looking at start and end point pallone and so what
we see with this data is that yes
our predictable patterns of selection but we also have a good abundance of
patterns that are not predictable and that kind of go against our expectations
of how genetic some of these markers should be changing in these groups so these patterns are pretty hard to
explain how that can be happening but there’s two ways that rationally make
sense of these dynamic patterns of change and the first is if we have the
marker near a gene that has contrasting total effects on larval fitness so in
this scenario gene a would have negative effects early on and then positive
effects later in the larval development period that would drive the frequency of
this marker down and then up another scenario is if it’s attached or if it’s
associated with two different genes with contrasting effects so in this case gene
B is the one driving it down and frequency early on but gene a drives it
back up yet so this is called repulsion phase so this could be one of the
scenarios that’s driving a lot of this behavior now the big takeaway here is
yes we saw 32% of all the markers that had significant changes were
bi-directional but it’s actually a bit more dramatic than that if you look at
just the start and end points you would only you would only conclude that a very
small proportion it actually changed when you account for these these
temporal changes in those intermediate time points we have a 63% that went
through some sort of balancing selection so that means that start and end points
are similar but if one of those ages in between we have significant changes so
this is evidence to suggest that there’s a lot more going on than would appear on
the surface just by looking at start and finish and this also leads us to perhaps
the conclusion that some of this is representative of balancing selection so
I’ll talk briefly about balancing selection typically in a population if
you have deleterious or negative alleles the traditional theory suggests that
those should be driven down in frequency through selection basically if you die
you lose that from a population one of the big mysteries with oysters is how
can they carry as much load as they seem to have and there’s a couple
explanations one is it appears they have a very high mutation rate so it’s been
estimated to be about 90 times greater than that of the fruit fly also these
populations have very non-traditional breeding characteristics they’re driven
by low to population sizes and Street pH
reproduction but this introduces another possible mechanism and if this balancing
selection is taking place it’s also another way where these negative alleles
are maintained in the population through this flip-flop pattern and that might
actually fold into our broader concept the success of this species so like we
said that if these species has found firm footing on all the continents
across the globe and demonstrated a significant ability to withstand
hazardous environments and this Center graph here is from the genome paper they
demonstrated they have a wealth of genes that have been specialized for a lot of
environmental stresses so perhaps in this case this balancing selection is
one of the mechanisms where they maintain their adaptability for future
volatile ocean scenarios so froy stirs perhaps because of this great fecundity
which is more or less unparalleled within the animal kingdom plus their
diversity this allows them to be adapt so kind of
these concepts up what we saw from our long-term phenotypic assays was that
long-term level response to a is complex it’s staged specific it seems to be much
more apparent that those more from morphological transitions whether that’s
the early development period or in that settlement phase and these represent
bottlenecks for that population where the stresses can kind of maybe be
manifest but there it’s also interrelated with a lot of other
components of that environment both biotic and abiotic which may dictate the
final outcome of that larval population and we also see a huge variation
year-to-year or between cohorts and so future work really should be focusing on
what are some of the variability to some of these effects rather than what’s just
the discrete phase that we can show repeated effects of Oh a but one of the
big takeaways for the vert shock program is that MEP did seem to have higher
performance across the board compared to wild stocks but it’s important to notice
to note that this isn’t hatchery condition so we have nothing to suggest
that these Fitness advantages will be maintained in natural environments where
the variation the environmentally action is much more substantial than looking at
the genetics we see that there’s a lot of genetic changes that are happening
just in larval development by itself that they attempt they tend to be with
broad physiological properties or physiological functions and there seems
to be some differences between those wild selected lines but interestingly
and encouragingly the effects of a way did we appear to be
much more consistent than just general developmental signal and specifically
membrane structure and performance appears to be a key component to the
survival of these animals in these hazardous environments so it’s
remember though that larval development is genetically chaotic so there’s a lot
of stuff that’s going on there that really hasn’t been investigated fully
yet and I think that these findings largely fortify some of the previous
estimates of genetic load and genetic mediated mortality and that genetics
might dominate a large portion of the mortality for this animal
and in this case balancing selection might make the purging of deleterious
genes difficult but also from a grading standpoint it might make the adoption of
new traits also difficult if you have this dynamic pattern of selection and a
period of life where they’re losing over 90 percent of their population but the
incur and the encouraging side MBP has shown improvement in field traits and
now we have some evidence to suggest improvement in larval traits and we also
have a permanent signal on the genetic changes over way on some of the spat so
it does appear that some of these effects can have long-term impacts for
better or worse on the oyster populations and I think I’m getting near
at a time but I’m happy to entertain any questions people have and if you have
any more long form questions I encourage you to shoot me an email I’m happy to
discuss any of this work with you down the road
great well thank you Evan for your presentation and your perspectives so if
anybody wants to type in a question for Evan into the question box over on your
right we’ll start the discussion don’t be shy
right now I don’t have any I think it’s because everybody is typing you have
Evans email it’s D URL a nd e VA n at and so if you have any
specific questions or would like a link to a paper or something that he
presented that would be a great place the other thing is is that we will put
this recording up on the California Current acidification that work website
as soon as all the processing is done on that on the behind-the-scenes portion so
it’ll be available for you to view at any time I know there are a number of
slides that I want to go back to and think about and look at
listen to your discussion again so any questions for Evan so I’m not seeing any pop up but that is
okay I made sure everybody can communicate um so you talked about
larvae being genetically chaotic you know having chaotic genetics you when
you did the mortality studies looking at the in the various vessels looking at
when ocean acidification or when they weren’t maturing was there any other
observations that you made relative to that really you know for future work or
for something as people are using these larvae in high and low Oh a conditions
that they should key in on as development occurs yeah one of the
things that we really well I wish we had more time to investigate was we actually
saw an increased growth rate or increased overall size of spat in high
co2 conditions but that is difficult to kind of parse whether that’s growth rate
or settlement rate and so I have a suspicion that actually the increased
size of speck in high co2 conditions might be driven by an accelerated
settlement rates meaning that they’re actually settling out earlier than their
counterparts and thereby entering that spat face sooner so that was one of my
big questions and that would kind of make sense as a stress response like
larvae that are in a know a condition decided we need to drop out of here and
that was interesting behavioral response from those analyst that’s I think
something that’s right for kind of looking further and I know some people
are looking at some of the effects of away on settlement all right well
something for another person working with the MVP lines to look at heat lynn
Willig has a question did any of your work look at the effect of diet Erna
l’ve aerations in co2 levels no and I think some of
that work is really fascinating because that’s kind of were environmentally
relevant but you know we we tried to maintain a stable of conditions as we
could get so we looked at this kind of one set level carried that through is
that something that perhaps could be done in the future
yeah and I know some people at Stony Point have been doing some diurnal work
and looking at kind of refuges and understanding how much exposure actually
start soliciting the response and frankly for us that was a system that
was difficult for us to set up but though I think that that’s very relevant
and important place to be taking some of this work trying to simulate this
natural environment more completely great so we have another question from
Alexis Valerie Orton I hope I got that right it seems like the wild oysters had
more genetic variation across the larval stages compared to the bird stocks but
there wasn’t a significant difference in survival under ocean acidification
conditions does this suggest that these variations aren’t contributing to
improve survival and since the brood stocks aren’t varying as much does this
indicate that they are pre adapted to increase co2 in terms of their baseline
membrane protein expressions yeah those are good observations so I can say for
one that this data doesn’t provide us enough information to to determine
overall genetic diversity I did look at that and see if we could use some of the
cool seek data and we don’t have a significant difference here but previous
work has demonstrated that as expected our selected lines have overall lower
genetic diversity than wild stocks but in this case we didn’t see that that was
correlated to any difference in Fitness I think that when you ask about baseline
membrane settings that’s a big question is why MVP stocks outperformed Wilde’s
and actually that’s part of the discussion one of our papers is that
really when you look at the data it’s that MBP did better overall period they
also did better in a way but they didn’t do especially better in fact their
fitness advantage is reduced so it’s my opinion and we don’t have data to to
kind of tease out some of those specifics but I think that MVP simply
after six generations of being reared in a hatchery is fine-tuned for those
conditions and can handle some of the other stresses may be a bit more easily
than while stocks
great any other questions for Evan she says God it thinks so with that I think
we will adjourn our webinar today and if you could advance the slide Evan so just
reminding everybody that this is recorded and that we will be presenting
it on the sea-can website if you have any in from you know if you’re
interested in future workshops or past workshops you can find them on the the
sea-can website you can also contract contact our coordinator Dianne pleasure
steel and her email is there and if we go to the next slide
our next roundtable will be a different date it’s August 7th 2019 at 1:00 p.m.
Pacific time and it is understanding acidification risks across habitats at a
ten site intertidal network and that is looking at eel grass and at Easter
culture in some specific plots it’s with dr. Micah Horwath who is with the
Washington Department of Natural Resources and the aquatic the aquatic
resource division and I think folks will find that webinar very interesting again
so thank you to dr. Derlin for taking the time to offer this his expertise and
perspectives about the genetics of larval fitness in the Pacific Way stir
all the way from Sweden for us and thank you all for taking the time to join us
this is the end of the session and we hope you join us again on August 7th
thank you

Leave a Reply

Your email address will not be published. Required fields are marked *

Back To Top