The evolution of trait correlations constrains phenotypic adaptation to high CO2 in a eukaryotic alga

Microbes form the base of food webs and drive biogeochemical cycling. Predicting the effects of microbial evolution on global elemental cycles remains a significant challenge due to the sheer number of interacting environmental and trait combinations. Here, we present an approach for integrating multivariate trait data into a predictive model of trait evolution. We investigated the outcome of thousands of possible adaptive walks parameterized using empirical evolution data from the alga Chlamydomonas exposed to high CO2. We found that the direction of historical bias (existing trait correlations) influenced both the rate of adaptation and the evolved phenotypes (trait combinations). Critically, we use fitness landscapes derived directly from empirical trait values to capture known evolutionary phenomena. This work demonstrates that ecological models need to represent both changes in traits and changes in the correlation between traits in order to accurately capture phytoplankton evolution and predict future shifts in elemental cycling.

1. Presentation of the model. I suggest adding more detail about what happens in each simulation, possibly in the supplement. What happens as it is stepped forward through generations? Are some individuals reproducing and others being lost according to fitness functions linked to current trait values? Or are they only experiencing changes in their trait values? (If this information was supplied and I missed it, my apologies.) I suspect these details are probably available in previous articles by the authors and others. However, I think it would be helpful to repeat it here, for two reasons. First, these details might bring assumptions that are worth stating explicitly. For example, the simulations probably assume steady state population dynamics, which presumably affect the fitness consequences of different traits, and potentially, combinations of traits. It seems like it would be important that the dynamics of the simulations match the data that inform their parameterization, or is this not necessary? I was particularly concerned about this because the results are discussed in the context of biogeochemical models, which often have regimes of both steady state conditions and highly episodic boom-bust growth conditions. Second, one of the traits that is examined is growth rate, which is an important parameter in many forms of microbial population models. I was initially confused that the growth rate data were being used to parameterize statistical models of trait change, rather than population dynamics of different lineages in a simulation. Possibly this will also confuse others that are more familiar with ODE models of microbe population dynamics, so I suggest it might be worthwhile to be very explicit about what is happening in the simulations.

Do you have any ethical concerns with this paper? No
Comments to the Author I'm not sure I fully understand the point of this paper -we know that trait correlations and initial phenotypes matter for evolutionary trajectories more generally -this is specific example of that but I'm not sure what we've learned generally form it. I don't have any problem with work per se, it's just the paper doesn't make it clear what new things we learn from this exercise. The conclusions are not different to what conclusions I would draw from reading say, Lynch and Walsh. I'm sorry I can't be more supportive, it's just I really don't know what is being argued here. The authors use a specific example, but then regard the traits themselves as generic, so I don't really see how the two things connect. If there's some deeper, novel point here, I'm sorry to say that it is beyond me.

15-Feb-2021
Dear Professor Levine: I am writing to inform you that your manuscript RSPB-2020-3144 entitled "The evolution of trait correlations constrains phenotypic adaptation to high CO2 in a eukaryotic alga" has, in its current form, been rejected for publication in Proceedings B.
This action has been taken on the advice of referees, who have recommended that substantial revisions are necessary. With this in mind we would be happy to consider a resubmission, provided the comments of the referees are fully addressed. However please note that this is not a provisional acceptance.
The resubmission will be treated as a new manuscript. However, we will approach the same reviewers if they are available and it is deemed appropriate to do so by the Editor. Please note that resubmissions must be submitted within six months of the date of this email. In exceptional circumstances, extensions may be possible if agreed with the Editorial Office. Manuscripts submitted after this date will be automatically rejected.
Please find below the comments made by the referees, not including confidential reports to the Editor, which I hope you will find useful. If you do choose to resubmit your manuscript, please upload the following: 1) A 'response to referees' document including details of how you have responded to the comments, and the adjustments you have made.
2) A clean copy of the manuscript and one with 'tracked changes' indicating your 'response to referees' comments document.
3) Line numbers in your main document. 4) Data -please see our policies on data sharing to ensure that you are complying (https://royalsociety.org/journals/authors/author-guidelines/#data).
To upload a resubmitted manuscript, log into http://mc.manuscriptcentral.com/prsb and enter your Author Centre, where you will find your manuscript title listed under "Manuscripts with Decisions." Under "Actions," click on "Create a Resubmission." Please be sure to indicate in your cover letter that it is a resubmission, and supply the previous reference number.

5
Sincerely, Dr Maurine Neiman mailto: proceedingsb@royalsociety.org Associate Editor Board Member: 1 Comments to Author: Thank you for submitting your manuscript to Proceedings B. Your manuscript has now been reviewed by two expert reviewers and myself. You will see that the reviews were mixed. Reviewer 2 did not see the novelty in the manuscript or how the findings presented could reasonably be generalized beyond the case presented. Reviewer 1 was more positive, finding the manuscript mostly well-written and potentially important, but also found it lacking in making a link between the specific simulation of trait variation and a more general inference about fitness within a physiological and biogeochemical context. Reviewer 1 also had criticisms about the description of the methods and choices made in designing the simulation. In summary, there is a disconnect between what the authors see as the implications and impact of this work and what is communicated by the manuscript.
Based on these reviews and my own evaluation, my recommendation is to reject with the option to resubmit. A revised manuscript would have to address each of the reviewer criticisms. The specific concerns about the presentation of the model and the evaluation of trait correlations raised by Reviewer 1 would need to be clearly and explicitly addressed. The more general concerns of both reviewers would also have to be satisfactorily dealt with: that the conclusions are not generalizable from this one case study, that the link between trait correlations and more general biogeochemical models is unsupported, and that therefore the work does not present a significant advance. Being mindful of the broad readership of Proceedings B, the narrative should be as accessible as possible to biologists and ecologists who are not modelers. If you choose to revise and resubmit, these changes should be reflected in the manuscript and in your point-bypoint response to the reviewers.
Reviewer(s)' Comments to Author: Referee: 1 Comments to the Author(s) This manuscript is concerned with multitrait variation and evolution. It examines ways in which evolving organisms might be constrained to access only a subset of available phenotype combinations in multitrait space. This is a worthwhile undertaking, concerned with processes that might limit the capacity of organisms to respond to change, and influence the biogeochemical cycles those organisms mediate.
The paper is excellent in most respects. The research outcomes are potentially interesting and important, and the manuscript is largely well-crafted. I had two comments, which are described in detail below. The first concerns the presentation of the model. I found it very hard to understand how the simulations worked, and recommend the addition of details, possibly to the supplement. The second is that I wondered about aspects of the rationale for evolving trait correlations, and wondered if this could be clarified.
1. Presentation of the model. I suggest adding more detail about what happens in each simulation, possibly in the supplement. What happens as it is stepped forward through generations? Are some individuals reproducing and others being lost according to fitness functions linked to current trait values? Or are they only experiencing changes in their trait values? (If this information was supplied and I missed it, my apologies.) I suspect these details are probably available in previous articles by the authors and others. However, I think it would be helpful to repeat it here, for two reasons. First, these details might bring assumptions that are worth stating explicitly. For example, the simulations probably assume steady state population dynamics, which presumably affect the fitness consequences of different traits, and potentially, combinations of traits. It seems like it would be important that the dynamics of the simulations match the data that inform their parameterization, or is this not necessary? I was particularly concerned about this because the results are discussed in the context of biogeochemical models, which often have regimes of both steady state conditions and highly episodic boom-bust growth conditions. Second, one of the traits that is examined is growth rate, which is an important parameter in many forms of microbial population models. I was initially confused that the growth rate data were being used to parameterize statistical models of trait change, rather than population dynamics of different lineages in a simulation. Possibly this will also confuse others that are more familiar with ODE models of microbe population dynamics, so I suggest it might be worthwhile to be very explicit about what is happening in the simulations.
2. The evolution of trait correlations. In the case of historical bias, I was curious about the rationale for imposing initial conditions of trait correlation values, but then allowing those values to take a random walk through the simulations. In broader applications (e.g. involving nutrient processing, temperature responses) key traits sometimes have strong tradeoffs. Is there a way to use the data to infer when there is justification for setting not only initial trait correlation values, but also constraining how much they can vary? Otherwise, over a sufficiently long simulation, it seems like organisms could evolve to have trait combinations that are functionally very unlikely (on a physical or chemical basis).
In sum, the article has the potential to make a useful contribution to the literature, and provides insights about how adaptation could be integrated into biogeochemical models. But the paper attempts to make quite a big conceptual link, between a statistical model for trait variation, and biogeochmical models where those traits could be used to inform condition-dependent parameters for transformations. Some of the contact points between those two things are not clear at present, and might be clarified by details of the simulations.
Referee: 2 Comments to the Author(s) I'm not sure I fully understand the point of this paper -we know that trait correlations and initial phenotypes matter for evolutionary trajectories more generally -this is specific example of that but I'm not sure what we've learned generally form it. I don't have any problem with work per se, it's just the paper doesn't make it clear what new things we learn from this exercise. The conclusions are not different to what conclusions I would draw from reading say, Lynch and Walsh. I'm sorry I can't be more supportive, it's just I really don't know what is being argued here. The authors use a specific example, but then regard the traits themselves as generic, so I don't really see how the two things connect. If there's some deeper, novel point here, I'm sorry to say that it is beyond me.

Do you have any concerns about statistical analyses in this paper? If so, please specify them explicitly in your report. No
It is a condition of publication that authors make their supporting data, code and materials available -either as supplementary material or hosted in an external repository. Please rate, if applicable, the supporting data on the following criteria.

Do you have any ethical concerns with this paper? No
Comments to the Author Based on the authors responses, I'm not convinced that this journal is the best venue for this work -the findings aren't novel to an evolution crowd and would seem to me to have much more potential for impact if aimed at a biogeochemistry crowd. But given the work was solicited from a preprint, I don't think this is my call to make. I have no specific comments about the ms as I think it's fine work.

07-May-2021
Dear Professor Levine I am pleased to inform you that your Review manuscript RSPB-2021-0940 entitled "The evolution of trait correlations constrains phenotypic adaptation to high CO<sub>2</sub> in a eukaryotic alga" has been accepted for publication in Proceedings B.
The referee(s) do not recommend any further changes. Therefore, please proof-read your manuscript carefully and upload your final files for publication. Because the schedule for publication is very tight, it is a condition of publication that you submit the revised version of your manuscript within 7 days. If you do not think you will be able to meet this date please let me know immediately.
To upload your manuscript, log into http://mc.manuscriptcentral.com/prsb and enter your Author Centre, where you will find your manuscript title listed under "Manuscripts with Decisions." Under "Actions," click on "Create a Revision." Your manuscript number has been appended to denote a revision.
You will be unable to make your revisions on the originally submitted version of the manuscript. Instead, upload a new version through your Author Centre.
Before uploading your revised files please make sure that you have: 1) A text file of the manuscript (doc, txt, rtf or tex), including the references, tables (including captions) and figure captions. Please remove any tracked changes from the text before submission. PDF files are not an accepted format for the "Main Document".
2) A separate electronic file of each figure (tiff, EPS or print-quality PDF preferred). The format should be produced directly from original creation package, or original software format. Please note that PowerPoint files are not accepted.
3) Electronic supplementary material: this should be contained in a separate file from the main text and the file name should contain the author's name and journal name, e.g authorname_procb_ESM_figures.pdf All supplementary materials accompanying an accepted article will be treated as in their final form. They will be published alongside the paper on the journal website and posted on the online figshare repository. Files on figshare will be made available approximately one week before the accompanying article so that the supplementary material can be attributed a unique DOI. Please see: https://royalsociety.org/journals/authors/author-guidelines/ 4) Data-Sharing and data citation It is a condition of publication that data supporting your paper are made available. Data should be made available either in the electronic supplementary material or through an appropriate repository. Details of how to access data should be included in your paper. Please see https://royalsociety.org/journals/ethics-policies/data-sharing-mining/ for more details.
If you wish to submit your data to Dryad (http://datadryad.org/) and have not already done so you can submit your data via this link http://datadryad.org/submit?journalID=RSPB&manu=RSPB-2021-0940 which will take you to your unique entry in the Dryad repository.
If you have already submitted your data to dryad you can make any necessary revisions to your dataset by following the above link. 5) For more information on our Licence to Publish, Open Access, Cover images and Media summaries, please visit https://royalsociety.org/journals/authors/author-guidelines/.
Once again, thank you for submitting your manuscript to Proceedings B and I look forward to receiving your final version. If you have any questions at all, please do not hesitate to get in touch.
Sincerely, Dr Maurine Neiman mailto:proceedingsb@royalsociety.org Associate Editor Comments to Author: Thank you for your revisions, which are comprehensive. There are no further revisions requested by the reviewers or associate editor.

Reviewer(s)' Comments to Author:
Referee: 2 Comments to the Author(s). Based on the authors responses, I'm not convinced that this journal is the best venue for this work -the findings aren't novel to an evolution crowd and would seem to me to have much more potential for impact if aimed at a biogeochemistry crowd. But given the work was solicited from a preprint, I don't think this is my call to make. I have no specific comments about the ms as I think it's fine work.

17-May-2021
Dear Professor Levine I am pleased to inform you that your manuscript entitled "The evolution of trait correlations constrains phenotypic adaptation to high CO2 in a eukaryotic alga" has been accepted for publication in Proceedings B.
You can expect to receive a proof of your article from our Production office in due course, please check your spam filter if you do not receive it. PLEASE NOTE: you will be given the exact page length of your paper which may be different from the estimation from Editorial and you may be asked to reduce your paper if it goes over the 10 page limit.
If you are likely to be away from e-mail contact please let us know. Due to rapid publication and an extremely tight schedule, if comments are not received, we may publish the paper as it stands.
If you have any queries regarding the production of your final article or the publication date please contact procb_proofs@royalsociety.org Your article has been estimated as being 10 pages long. Our Production Office will be able to confirm the exact length at proof stage.
Data Accessibility section Please remember to make any data sets live prior to publication, and update any links as needed when you receive a proof to check. It is good practice to also add data sets to your reference list.
Open Access You are invited to opt for Open Access, making your freely available to all as soon as it is ready for publication under a CCBY licence. Our article processing charge for Open Access is £1700. Corresponding authors from member institutions (http://royalsocietypublishing.org/site/librarians/allmembers.xhtml) receive a 25% discount to these charges. For more information please visit http://royalsocietypublishing.org/open-access.
Paper charges An e-mail request for payment of any related charges will be sent out shortly. The preferred payment method is by credit card; however, other payment options are available.
Electronic supplementary material: All supplementary materials accompanying an accepted article will be treated as in their final form. They will be published alongside the paper on the journal website and posted on the online figshare repository. Files on figshare will be made available approximately one week before the accompanying article so that the supplementary material can be attributed a unique DOI.
You are allowed to post any version of your manuscript on a personal website, repository or preprint server. However, the work remains under media embargo and you should not discuss it with the press until the date of publication. Please visit https://royalsociety.org/journals/ethicspolicies/media-embargo for more information.
Thank you for your fine contribution. On behalf of the Editors of the Proceedings B, we look forward to your continued contributions to the Journal. We are resubmitting our manuscript titled "The evolution of trait correlations constrains phenotypic adaptation to high CO2 in a eukaryotic alga" for publication as an Article in Proceedings of the Royal Society B. The original submission (RSPB-2020-3144) was solicited by Dr. Maurine Neiman of the Proceedings B Preprint Editorial Team who found our paper on BioRxiv and encouraged us to submit to Proc B. We have carefully noted your comments and suggestions, and the Reviewers' comments and suggestions, in producing this revised version of the manuscript. We think these comments have helped to further clarify our results and strengthen the paper. Specifically, we have substantially revised the main text, including rewriting the entire discussion, to highlight the novelty of our work. We have conducted a new model analysis to demonstrate the ability of our approach to bridge evolutionary models and ODE-based biogeochemical models. Finally, we have run additional sensitivity analyses and expanded the description of our model. We detail these changes in the point-by-point response to the Reviewers' comments.
Microbes form the base of global biosphere and drive both aquatic and terrestrial biogeochemical cycling, thereby significantly influencing Earth's food webs and climate. Predicting how microbial populations adapt and how this will influence global elemental cycles remains a fundamental challenge. We are at a critical juncture where we need ecosystem and biogeochemical models to incorporate evolutionary dynamics in order to robustly predict future shifts in ecosystem dynamics. The primary aim of this paper was to create a bridge between these two disciplines by developing a novel framework based on easily quantifiable microbial traits. We show that this new framework is consistent with existing evolutionary theory and can be used to understand constraints on quantifiable trait-based phenotypes such as those measured in the laboratory and field and modeled by biogeochemists. We thus have provided a framework which can be used with easily obtained trait data to understand constraints on phenotype adaptation for globally relevant microbes. An additional major advance in our approach is that it can be used on data from unculturable organisms and is scalable to a large number of traits. Our work further demonstrates that ecological models need to represent both changes in traits (already existing in some ecological models) and changes in the correlation between traits in order to accurately capture phytoplankton evolutionand proposes a framework for such an integration. We believe our findings are of interest to widespread scientific fields including evolutionary biology, physiology, biological and chemical oceanography, aquatic ecology, biogeochemistry, and ecosystem modeling.
The We assume a constant population size (1000 individuals) over the entire simulation. We be overcome (ie -when Ne was not extremely small).

36
As TRACE uses an adaptive random walk to capture evolutionary dynamics, we are not 37 directly simulating population dynamics the way an ODE based ecosystem model would. A more 38 detailed comparison between our model and typical ODE ecosystems models is provided below in 39 response to comment 1c. 40 41 1c. I was particularly concerned about this because the results are discussed in the context of 42 biogeochemical models, which often have regimes of both steady state conditions and highly 43 episodic boom-bust growth conditions. Second, one of the traits that is examined is growth rate, 44 which is an important parameter in many forms of microbial population models. I was initially 45 confused that the growth rate data were being used to parameterize statistical models of trait 46 change, rather than population dynamics of different lineages in a simulation. Possibly this will 47 also confuse others that are more familiar with ODE models of microbe population dynamics, so 48

I suggest it might be worthwhile to be very explicit about what is happening in the simulations. 49
Response: This highlights an important point which we have tried to clarify in the revised text. 50 The TRACE model presented in this paper captures shifts in traits and trait-correlations that occur 51 as phytoplankton adapt to a new environment. TRACE is a novel framework which allows us to 52 link evolutionary theory of how adaptation proceeds with quantifiable microbial traits. In this 53 paper, we demonstrate that correlations between traits in the ancestral populations (akin to 54 historical effects on adaptation, or genetic constraints) constrain the adaptive walk of the 55 population and influence the evolved phenotypes. To ground our model in observations, we use 56 observed shifts in traits and trait relationship (quantified using PCA) from an experimental 57 evolution study with Chlamydomonas. Growth rate was one of the traits quantified in the 58 experiment which not only shifted significantly between the ancestral and evolved populations but 59 also changed in relationship to other quantified traits. As such, we selected it as one of the traits to 60 simulate in TRACE -however, TRACE is agnostic as to what the actual traits are. Because the 61 dynamics within TRACE are not linked to the growth dynamics of the individuals, including 62 growth rate as a trait is justified. We agree that this could create some confusion and so have tried 63 to clarify this point in the text and refer to the traits simply as traits 1-4 to help minimize confusion. 64 An additional reason for selecting growth rate as a trait of interest is that it is a commonly 65 measured trait in both laboratory and field based studies. As such, we felt including growth rate 66 might resonate with other researchers who might want to use TRACE to better understand their 67 system. Since growth rate is typically used as a key proxy for fitness in microbial populations, we 68 believe it is important to examine how it relates to other ecologically important traits in a 69 multivariate landscape, or the integrated phenotype. However, it is worth noting that, in the context 70 of an evolutionary model, fitness is a relative value. Specifically, fitness is the relative reproductive One of the key challenges in understanding how phytoplankton will adapt to shifts in the 76 environments is that, at present, evolutionary models are not suited to interface with 77 biogeochemical ODE models (Ward et al. 2019). One of our aims in developing TRACE was to 78 bring evolutionary models and trait-based ecosystems models (ODE models) closer together. 79 While TRACE is not a mechanistic model (e.g. growth rate is not solved for prognostically), it 80 provides important insight into how trait adaptation might be implemented in ODE models. in the trait-scape closer to the optimum. However, because trait 1 is linked to the other traits, the 128 change in trait 1 (which is identical in individual A and B) will result in changes to traits 2-N that 129 will not be identical in individual A and B because of the different trait-correlations. As a result, 130 A and B will not have identical movements in the trait-scape and one most likely will be closer to 131 the high-fitness optimum and therefore more likely to persist. What is exciting about our approach, 132 is that it provides insight into how the trait-correlations present in the initial population impact the 133 adaptive outcome. This highlights the importance of including both changes in trait values and 134 changes in trait-correlations into marine ecosystem models that represent adaptation. Furthermore, 135 because TRACE is parameterized using high-fitness trait combinations derived from actual 136 observations, there is not a concern that the model will generate organisms with trait-combinations 137 that are not physically or chemically possible. We have modified the discussion to highlight this 138 point and have added in additional simulations using an ODE based ecosystem model to reinforce 139 this point (see above). 140 The reviewer's comment brings up an additional point -while we vary the 'genetic diversity' 141 in our starting population (variable trait-correlation), we chose to seed the model with a single 142 phenotype (trait values). Specifically, we selected the starting point based on the observed trait 143 values from the experimental evolution study. It is important to note that while all individuals 144 started with the same phenotype, they were not a true clonal population in the model as the trait-145 correlations varied between individuals for most of our simulations. However, the reviewer brings 146 up an important point that relates to one of the key findings of our work -often there are multiple 147 high-fitness trait combinations for a given environment. To address this, we have conducted an 148 additional set of simulations in which we initialize the model with a diverse population both in 149 terms of trait-correlations and in terms of trait values (phenotypes). To link these additional 150 simulations with the empirical data, we use 4 different phenotypes that were present in the ancestral 151 data of the Lindberg & Collins (2020) experiment. These runs converged on the same 4 emergent 152 phenotypes as in Fig. 4 indicating that the phenotypic diversity of the initial population does not 153 impact our results. Due to space constraints, the figures for these additional simulations are 154 included in the supplement (Supplementary Fig. 1) but are referred to on lines 269 -273 of the 155 revised text. 156 157 158 trait data can easily be collected for both cultured and uncultured representatives such that this 250 approach can have wide-reaching utility. For example, while we focus on a case study using an 251 experimental evolution study to generate the trait-scape, our framework can be used to generate a 252 trait-scape for a given environment A using data from an in situ population (assuming that the 253 population is well adapted to the environment they are living in). TRACE could then be used to 254 develop hypotheses as to how a new population from a different environment B might adapt upon 255 exposure to environment A. 256 257