[This is a review of the original version. See Text S1 for the version history. The authors’ responses are included in line and are reflected in the published version.]

Summary

I think a wiki page on this topic is timely and needed. Part of the reason is that the digital evolution field utilizes environments/jargon that are likely to be just as foreign to a molecular biologist as molecular biology is to a classically-trained computer scientist. In addition, the digital evolution field has seen a number of major developments, coinciding with an increased interest in experimental evolution. However, the current topic page has a long way to go to try and bridge this gap. If it does, it could substantially help communication and cross-talk between computer scientists and evolutionary biologists. I hope my comments are helpful in trying to reach this goal. In revising this page, please consider the following questions:

Response: We appreciate very much the positive and constructive comments made by the reviewer. His concerns about this topic page are well supported and we have tried to take advantage of them for improving the page. We hope, as he certainly pushes for advance, that this topic page helps advance communication and cross-talk between computer scientists and evolutionary biologists. That is our ultimate goal. Indeed, these new topic pages by PLoS Computational Biology seem like a great step towards the sort of community required to bridge this gap, where authors can seed “living” articles on wikipedia.

1. Is this topic page meant to enable a biologist to implement and/or understand how a particular digital model of coevolution works? Or, to help a CS person build digital evolution models? Or both? I think the objective needs to be clearer from the outset.

Response: The goal is to make a bridge between these two research fields. This is a collaboration between an ecologist and computer scientists to introduce a new empirical tool for studying the evolution of species interaction networks in artificial evolving systems. The key term here is interaction networks, not only digital evolution (which has been studied for the last 15 years). To make this objective clearer, we have reorganized the sections of the topic page and explicitly highlighted this point in the new section entitled "Inspiration": "This topic page highlights new research avenues enabled by the inclusion of ecological interactions in digital systems. Research using self-replicating computer programs complements laboratory efforts by broadening the breadth of viable experiments focused on the emergence and diversification of coevolving interactions in complex communities. This cross-disciplinary research program provides fertile grounds for new collaborations between computer scientists and evolutionary biologists".

2. The article focuses on digital models of a certain kind, but does not discuss coevolutionary studies based on ODE models, many of which have internal states, e.g. see the work of Childs et al. (Evolution, 2012) (disclosure: I am a corresponding author on that study) and Weinberger et al. (PLoS Comp Biol, 2012) on coevolutionary dynamics induced by a newly discovered form of adaptive immunity in bacteria/archaea called “CRISPR/Cas immune defense”. Do these models fit the definition? I don’t think the authors necessarily need to include them, however my point is that the article should be clearer on its scope and what is meant by “computer programs”.

Response: We have tried to clarify this point in the definition of "evolving digital ecological networks", at the beginning of the topic page, and in the “Digital organism” section. The novelty of our approach is to explore the networks of ecological interactions between digital organisms, i.e. "self-replicating computer programs that evolve in an open-ended way", as they have been implemented and defined in artificial life software platforms like Avida. The goal of this topic page is not the introduction of a particular model of coevolution (that, as the reviewer said, can be studied using mathematical models based on ODEs), but a research platform for collecting data for studying different mechanisms of coevolution in a community context. To make this point clear, we have expanded the introductory section and the description of the interactions between digital organisms in three subsections entitled "Host-parasite interactions", "Mutualistic interactions", and "Predator-prey interactions".

3. What is the point of the research section? It is largely speculative, yet many articles have been written on AVIDA/Tierra/etc. Are these research directions or conclusions from research?

Response: The reviewer is definitely right. This question was also raised by the other reviewer. Since this is not a research paper, our aim is to highlight some open questions that can be addressed by using this approach (some of which we are already working on). Following the reviewer's suggestion, this section is now entitled "Research directions" (see also our related reply in the "Minor comments").

Major comments

1. I don’t think it is necessary to criticize experimental evolution studies as being too hard, too slow, etc. as a means to justify the use of digital coevolution studies. In fact such criticism is counter-productive. The reality is that Richard Lenski, his colleagues, and their peers, have been developing experimental methods to test biological evolution for decades. These methods are becoming increasingly standard, and at least for microbes, testing evolution is no longer a pie-in-the-sky idea. I think greater caution and respect to the innovations of the experimentalists is in order here. The section in the Introduction should be recast substantially. The point to be made is to emphasize the positive contribution that models can make in this area, because ultimately, biologists are interested in how evolution takes place in “real” systems, and not within the “digital world” (I say this knowing full well that there are an increasingly number of “real digital worlds” for which these studies may also be relevant).

Response: We are aware that we have been overly critical of the literature on experimental biological evolution. This was due to our enthusiasm for the novel incorporation of ecology into digital evolutionary studies of communities of interacting organisms. Our goal does not focus on testing evolution (for which there is a rich body of knowledge), but on incorporating ecology in to experimental evolution studies. Indeed, we know the excellent work carried out by Richard Lenski and his team very well (the second author of this topic page is doing part of his PhD research in Lenski’s lab). However, the reviewer is certainly right when suggesting that our criticism is counter-productive. We have taken his advice and improved the introduction to emphasize the contribution that the digital evolution of interactions can make in this area (see our reply to this in the "Minor comments" section below).

2. It is true that digital environments do not include the “limitations evolutionary biologists typically face”. However, utilizing digital environments comes with a substantial caveat: do the microscale rules influence the types of structures that evolutioanr biologists would like to understand? If so, how then does the choice of “substrate” i.e., logic gates vs. nucleic acids, affect predictions? I believe an additional section needs to be included on caveats in translating predictions of evolving digital communities to biological systems. Otherwise the claims in the last paragraph of the Introduction may seem overly optimistic: “Evolving digital organisms within ecological networks can help to elucidate how selection on one species affects the evolution of links among the other species throughout the web”…. Why might this not be the case? The lack of text on this topic will not be reassuring to a reader unfamiliar with this literature.

Response: Although we are looking for general principles that can help understand what happens in ecological communities, it is true that we should include potential caveats when translating predictions of evolving digital webs to biological systems. In order to address these limitations we have incorporated the following paragraph at the end of the topic page: "The study of self-replicating and evolving computer programs offers a tantalizing glimpse into the evolution of interactions among organisms that do not share any ancestry with biochemical life of Earth. This comes with potential caveats in translating predictions of evolving digital networks to biological ones because mechanistic details differ substantially between interacting digital organisms and interacting biological organisms. Nevertheless, these digital networks contain the necessary components for ongoing coevolutionary dynamics in large webs of interacting organisms. In spite of the differences between biological and digital evolution, the study of evolving digital ecological networks can lead to a more predictive understanding of natural dynamics. Because the general operational processes (e.g., Darwinian evolution, mutualism, parasitism, etc.) do not differ, studies utilizing digital networks will uncover rules operating on and within ecological networks. Together with microbial experiments, they create opportunities for furthering our understanding of the interplay between ecological and evolutionary processes among interacting species".

Minor comments:

Stylistic
• References seem to be missing/badly hyperlinked (i.e., lots of [wp:…] type of elements in the text)

Response: We apologize for the formatting of the article upon review. We were likely in the process of adding more wikipedia links to the web version while it was being reviewed. We have fixed the links that were badly hyperlinked: patterns of interactions, simulation, stacks, insertions, deletions, and ecological networks. The use of hyperlinked elements was a requirement of the journal for the Topic Pages since "at the point of publication, the Topic Page will be added to Wikipedia, becoming a living page that can be edited and updated by the community".

Introduction
• are not evolving -> do not evolve

Response: Corrected.

• Figure 1: species coevolving within large networks (shouldn’t there be a citation to the “diffuse coevolution” work from the 1980s and 1990s on these ideas)

Response: The reviewer is correct, we should have referenced the "diffuse coevolution" ideas that Janzen originally proposed. We do want to be careful not to misinterpret diffuse coevolution to mean the confusion of pair-wise adaptation between interacting species, and rather that it is an additional level at which selective pressures may occur. In other words, we want to stress the importance of thinking about a network context where specific interactions between species exist and how those interactions shape coevolutionary pressures. The following sentences have been added to the introductory section: "The concept of diffuse coevolution, where adaptation is in response to a suite of biotic interactions, was the first step towards a framework unifying relevant theories in community ecology and coevolution. Understanding how individual interactions within networks influence coevolution, and conversely how coevolution influences the overall structure of networks, requires an appreciation for how pair-wise interactions change due to their broader community contexts as well as how this community context shapes selective pressures. Accordingly, research is now focusing on how reciprocal selection influences and is embedded within the structure of multispecies interactive webs, not only on particular species in isolation".

• Figure 3 is referenced in text before Figure 2 is.

Response: We have corrected this by interchanging the order of the figures to follow the current, reorganized flow of the topic page.

• The examples section needs to take a step back, I feel I am midway through an article. This is the first example of what a digital evolving organism is. Please define and go slowly here.

Response: It is true that the previous organization of the topic page was confusing (in part because the topic pages template is quite different from an standard research paper). We have now removed that section and expanded the “Implementation” section, including a better explanation of what a digital organism is and the types of interactions implemented so far.

• Figure 2 needs to be improved: where is the digital organism? You talk about in the legend, but it is not labeled in the Figure. I would suggest more of a multi-scale figure, step back and draw circles representing digital organisms, and then let us look inside to see what they’re made of, and then show how things change. So much is happening in this Figure that I’m afraid only those who already know this field will understand it.

Response: We believe this was a problem caused by the previous order of the figures. Figure 2 is now Figure 3. The new order allows the reader to grasp what a digital organism is and how the self-replication process takes place. With this change, we believe the reader is prepared to understand how the digital organism interacts with the environment by performing logical computations (tasks). The main message of the figure is not to illustrate what a digital organism is made of (which can be seen in what is now Figure 2), but to show that it can express different phenotypes depending on the tasks it performs. We believe this is the key figure of our topic page. Although we recognize that it condenses a lot of information, we think it is better to have all the panels in the same figure in order to facilitate the understanding of the piece as a whole. However, if the reviewer still feels it is better to split this figure in multiple parts, we will be willing to do so, since it depicts key elements of evolving digital ecological networks.

History
• The “Adding ecology” section is highly jargonized. For example: “Phenotypic convergence in species within the same trophic level and trait complementarity in interacting partners have been recently identified as the outcome of the coevolutionary process in mutualistic networks using individual-based simulations.” What does this mean???

Response: We agree with the reviewer about the organization and terminology used in this section. We have reorganized the topic page to improve the rationale behind the way interactions between digital organisms are implemented in this artificial life software platform, including changing the highlighted phrases pointed out by the reviewer. We have removed the section "Adding ecology" and expanded the description of the interactions between digital organisms in three different subsections entitled “Host-parasite interactions”, "Mutualistic interactions", and "Predator-prey interactions", all of them located within the section "Digital interactions".

• Tierra has no wiki page. Avida has a minimal page, and yet the current history does not really fill in the gaps as well. What is the plan to bulk this up?

Response: We would like to thank the reviewer for pointing out our lack of a hyperlink to the Tierra wiki page, http://en.wikipedia.org/w...(computer_simulation), and it is now referenced and linked in the main text ("Coreworld subsection"). By following an external link within that wiki page, the Tierra home page (http://life.ou.edu/tierra...) can be reached. The later has quite a lot information about this artificial life software platform. However, it is true that Avida has a minimal page in Wikipedia (http://en.wikipedia.org/w...). But, from that wiki page there is an external link pointing to the Avida download page (http://avida.devosoft.org...). Nevertheless, in order to facilitate the reader's access to the Avida research platform, we have included the following sentence in the "Avida" subsection: "The Devolab at the BEACON Center currently continues development of Avida". While we agree that the history section is short, we do not think that this topic page is the appropriate outlet for a broader treatment of the digital life’s history. Instead we aimed to provide enough context for those interested in the history of the field to find further information.

Implementation
• Figure 3: “the experimenter” -> “the modeler”.

Response: We don’t agree with the reviewer here. Avida is a experimental research tool in which researchers design experiments, not models.

•Figure 3: Also there are many undefined terms here. Does a non-CS person need to know these details? What are the letters? Why are there two circles? The animation seems beautiful, but I would strongly suggest a voiceover to link the animation and the caption. The animation lasts for 30 seconds, but just reading the caption quickly takes longer than that. I would suggest a voice-over (or closed-caption text) that walks the reader slowly through this Figure. This is a key Figure and yet right now, for non-experts, the mental leap between this figure and the caption will be difficult to say the least.

Response: In order to have a smooth flow while keeping the reader engaged, Figure 3 is now Figure 2. We have expanded the text in the figure legend in order to improve the clarity of what a digital organism is. We explicitly made reference to what is located on the right and on the left: "The circular genome of a digital organism, on the left, consists of a set of instructions (represented here as letters). Some of these instructions are involved in the copy process and others in completing computational tasks. The experimenter determines the probability of mutations. Copy mutations occur when an instruction is copied incorrectly, and is instead replaced by a random instruction in the forming offspring's genome (as can be seen in the offspring, on the right)". It is true that reading the caption takes longer than watching the animation. However, we honestly believe that the main message of the figure can be understood with a quick look at it, since it only represents the set of instructions a digital organism is made of and how mutations can occur during the copy process. Explaining what each instruction means will reduce the readability of the topic page. The reference and the direct link to Avida-ED is given at the end of the legend so that anyone interested in more details can easily get them. We agree that the description of the three pointers is too detailed for life scientists, but we decided to explain them since they appear during the animation. If we are not wrong, the animation will appear only in the wiki page, while in the printed topic page this figure will show the last frame of the animation. That is, it will depict the two digital organisms (the parent on the left and the offspring on the right) at the end of the self-replication process. For the printed version the following sentences (those we think are pointed out by the reviewer as too detailed and unnecessary) can be removed, reducing the time needed for reading the figure legend: "Initially, all three of the parent's hardware pointers are in the same location, at the instruction represented here by r. As execution begins, the instruction pointer (indicated by an i) advances. The first few instructions allocate space for the offspring, and then move the write-head pointer (indicated by a w) into that space. The flow pointer (indicated by an f) is used to move the other pointers to genetically specified locations. The remainder of the process of self replication is carried out by a set of instructions at the end of the genome, commonly referred to as the copy-loop. When execution reaches the copy-loop, the flow pointer is used to keep the flow of execution inside of a loop that advances the read and write heads and copies instructions from the parent genome (read-head) to the offspring genome (write-head). Arcs inside the circular genome represent the execution flow, showing most of the CPU cycles being used during the copying process".

• 32-bit? Really? Is 64-bit or 16-bit not allowed?

Response: The reviewer was correct in pointing out that this was an unnecessary detail to include. The reference has since been removed. But, to answer the reviewers question: any number of bits may be used, but 16 is short enough to occasionally trigger false-positives during evolutionary runs, 32 is sufficient not to, and thus 64 is overkill.

• The reference to nand is obscure. I assume it’s because nand is a universal gate, but do others? Shouldn’t you say this?

Response: We chose to mention NAND since it the only logic function included in Avida’s genetic code by default, and computer scientists may be interested in this particular detail. We have added more context and mentioned that NAND is a universal logic function to this section: "Digital organisms can acquire random binary numbers from the environment and are able to manipulate them using their genetic instructions, including the logic instruction NAND. With only this instruction, digital organisms can compute any other task by stringing together various operations because NAND is a universal logic function".

• Figure 4 animation does not work on my browser.

Response: The technical editor for Topic Pages (Daniel Mietchen) will help on this. If we are right, the animation will appear only in the wiki page. We believe that the figure legend is clear enough, at it is now, even if it depicts the last frame of the animation. We hope the reviewer can watch the animation now, or at least, one of its frames.

Research
• Please do not criticize experimentalists unnecessarily as a justification for your argument, e.g.: “but those data are completely lacking for natural systems.” Let me point out a compilation of phage-bacteria interaction networks, including evolutionary studies, by my group: Flores et al. PNAS (2011). Obviously a complete evolutionary history in each study is not available, but the idea that the data is “completely lacking” overstates the case: http://www.pnas.org/conte.... In addition Justin Meyer (who is acknowledged on this page) has done experimental coevolution between phages and bacteria, including sequencing and found very intriguing patterns of how coevolution takes place. Again, it’s not everything, but it’s not “completely lacking” either.

Response: As we said above, we have been overly critical of the literature on experimental biological evolution. We consider it a mistake to have excluded the meta-analysis of laboratory studies carried out by Weitz and his team, since that paper studies host-phage interactions across a broad spectrum of taxa, habitat and mode selection. This is a very good review of the empirical biological evolutionary studies of interacting taxa in a community context. The nested structure found and the absence of evidence for modularity is, to our knowledge, the first step towards investigating the interplay between evolutionary processes and network structure in laboratory experiments. Recognition of that paper is clearly needed since it lays at the heart of the main goal of this topic page. Accordingly, we have commented on the main messages of that paper in our introductory section: "Coevolution in a community context can be addressed theoretically via mathematical modeling and simulation, by looking at ancient footprints of evolutionary history via ecological patterns that persist and are observable today, and by performing laboratory experiments with microorganisms. In spite of the long time scales involved and the substantial effort that is necessary to isolate and quantify samples, the latter approach of testing biological evolution in the lab has been successful over the last two decades. However, studying the evolution of interspecific interactions, which involves dealing with more complex webs of multiple interacting species, has proven to be a much more difficult challenge. The meta-analysis of host-phage interactions in a broad spectrum of taxa and habitats, carried out by Flores and his team, represents one of the first steps towards experimental laboratory investigation of the interplay between evolutionary processes and network structure in biological systems". We have also included a reference to Justin Meyer’s paper, which was published after our original submission: "Studying the evolution of species interaction networks in these artificial evolving systems also contributes to the development of the field while overcoming limitations evolutionary biologists may face. For example, laboratory studies have shown that historical contingency can enable or impede the outcome of the interactions between bacteria and phage depending on the order in which mutations occur: the phage often, but not always, evolve the ability to infect a novel host". We thank the reviewer for reminding us to add this before finalizing it.

• This section is rather short, what is the objective?

Response: Since the goal of the Topic Pages is to introduce a new computational method to a broad audience emphasizing its potential as a research tool, we have presented some open questions that can be addressed by using this approach here. Following the reviewer suggestion (see his "Summary"), this section is now entitled "Research directions". As was suggested by the second reviewer, we have now added a very interesting question regarding the diversity-stability debate when different types of interactions are taken into account.