All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
The issues raised by the reviewers have been carefully considered and addressed.
[# PeerJ Staff Note - this decision was reviewed and approved by Valeria Souza, a PeerJ Section Editor covering this Section #]
All three reviewers raise a series of important issues about the narrative, missing details on the methods, experimental design, and framing of the research question, along with several minor corrections. These shortcomings should be carefully addressed to improve the paper.
[# PeerJ Staff Note: Please ensure that all review comments are addressed in a rebuttal letter and any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate. It is a common mistake to address reviewer questions in the rebuttal letter but not in the revised manuscript. If a reviewer raised a question then your readers will probably have the same question so you should ensure that the manuscript can stand alone without the rebuttal letter. Directions on how to prepare a rebuttal letter can be found at: https://peerj.com/benefits/academic-rebuttal-letters/ #]
no comment
no comment
no comment
GENERAL COMMENTS
The main challenge I found with understanding this manuscript was to follow the narrative. There appear to have been multiple apporaches taken, at various times, for various reasons. For example, in the Experimental evolution section, lines 459-472, there was a 25-day evolution experiment (two barcoded per well), and also mention of "532 longitudinal-evolutionary dynamics samples". I was confused. The dicsssino is much clearer, and does not need to change.
The Strains, media and culture methods section P9, line 109- do not explain well what the 92 diploid barcoded strains are.
For example, are they geetically different from noe another? Have they been sequenced? How heterozygous was the parent YPS163?
How different are they genetically from the ‘reference’ strain (barcode ID: d1H10) This is also the case in the results: P23, line 437.
STill at this point I don't know what the strains are. It was only in the discussion that I was sure that the barcoded strains started as clocal replicates.
hHis need to be explciit earlier, as it is critical to any readers understanding the experiment.
The 143 Experimental design section (P10, lines 143-) could be clearer.
Even with Figure 1 (which is good), I was not always clear about how many strains per well.
Perhaps show this to a colleage (that was not inolved) and check that they can understand samples, pools etc.
Data: there are no issues with the availabily of the raw data, or the code. All these are aspects are at or above the current standard.
Data analysis is also excellent. Sequence analysis and laboratory cuture also appear rigorous.
P24, line 453:
1% fitness differences does not seem very high, given that new alleles typically have effect sizes much smaller. How does this compare to other methods?
PS: Please provide figures in line with the text in future. Not at the end. Placing figures at the end is counter-productive as it makes the reviewers job harder. The journal should insist on this, as well (I know that most don't).
P24, line 453 "(n=22), we estimate high power (99.8%) to detect 1% fitness differences between treatments at a"
You estimate power for 22 strains, at this point in the ms. Why did you tehn conduct the 250 day experiments with two barcodes per well?
This should be explained in the ms.
SMALL CORRECTIONS
All page numbers ref to the page in the pdf.
P4, "fitness gains (adaptation) in a total of 76 experimentally evolving populations by conducting 1,216 fitness"
It would be useful to define what kind of populations here. Mitotic only or with meiosis? Starting from clonal, or a genetically heterogenous mix?
P7, line 69 "10,26,27]. However, the number of fitness assays, and thereby the resolution of those assays,"
reference style differs
P8, line 90: "Saccharomyces cerevisiae (hereafter, yeast) strains,"
At an early stage, please tell the reader what the strains are. A gene knockouts? A population of wild strains? etc.
Otherwise all the futher reading I am wonder what the experiments are doing to show.
(at the this point in the ms I am already convinced that bar codes are a good idea)
The Strains, media and culture methods section P9, line 109- still does not explain well what the strains in use are.
This is critical to understandin the experiment.
P9: line 134
"Saturated 24-hour culture was diluted (1:1000) into fresh medium at
134 the same time each day to initialize the next round of growth for all evolution and fitness assays."
What was the OD at this point? Did they reach stationary phase?
P13, line 201
"20, 22, 24 and 25) for a total of 532 evolutionary dynamics samples across 76 evolving two-"
Adding numbers such as "532 evolutionary dynamics samples across 76" does not help to clarify the experiment.
P11, line 170
Why 152 barcoded yeast strains used here?. At each point, please explain why certain numbers of strauis were chosen.
There are many sentences descrbing how many fitness assays can be conducted, for example:
"With four replicate fitness assays for each barcode, this amounted to 1,216 fitness assays (2 barcodes * 2 time-points 468 * 76 populations * 4 replicates)." P 24, line 267. And again P24, line 467, "fitness assays for each barcode, this amounted to 1,216 fitness assays (2 barcodes * 2 time-points".
I wonder how many of these statements help the reader to understand the experiment? We all apreciate, from an eary stage in them manuscipt, that the system allows for many repeats. That bar-seq is a cheaper option that full genome sequencing is well-known.
I felt that the article came a little short in terms of the background/context provided, specifically with reference to how this methods fits in with other similar methods. The authors should more clearly describe the similarities and differences. Please see more comments below.
Experimental design is sound.
The findings are supported by the data.
The manuscript “High-throughput analysis of adaptation using barcoded strains of Saccharomyces cerevisiae” presents a system to both measure fitness and conduct evolution experiments in budding yeast. The authors perform detailed analysis to assess the power of their system for detecting small differences in fitness. They also describe new experimental results confirming observations seen in previous studies, for example, that the rate of adaptation is faster in haploids than diploids. All in all, this is a nice study. My two major comments turned out to be very related, and so I really only have one main piece of advice that I hope improves the paper.
Major comments:
1. Another similar barcode system (the Levy/Blundell system cited by this study) achieves many of the same goals and the authors should take the time to describe all of the similarities and differences with that system. Here are some examples:
a. These systems have barcodes inserted into different regions of the genome. Is there any benefit to inserting into HO? Does the observation that barcode systems using a different barcode insertion locus are successful create possibilities for future multiplexing?
b. The previous system measures fitness by finding the log-linear slope of barcode frequency change across multiple timepoints (see cited Venkataram et al reference from 2016). The system in this paper uses only the start and end point, which has been shown to yield reduced power in the following reference: https://www.cell.com/cell-systems/pdfExtended/S2405-4712(18)30390-9. Is there a future plan to gain power by considering more timepoints in the fitness measurements?
c. The previous system does not pair the barcodes but combines hundreds of thousands. Is that something that you can do with your system? Is the max throughput of your system presumably similar to that seen in previous studies? Why did you choose to pair barcodes in the evolutions rather than combine hundreds of barcodes?
d. Would having performed a two-step PCR, as in the previous system, have decreased noise due to PCR jackpotting? Is this going to become a requirement if the current system is used for evolutions tracking pools of thousands of barcodes?
2. The authors argue that their system increases the throughput of fitness measurements and experimental evolutions. But a previous system (see cited Venkataram et al study 2016 and Levy & Blundell et al 2015) is actually much higher-throughput. The authors should be clear about this by making the following changes:
a. In the beginning of the conclusion, instead of saying that this system improves throughput, perhaps add a qualifier, something like, “improves throughput relative to other evolution experiments using paired strains with fluorescent markers”.
b. Are there specific questions that can be addressed with this system that are less amenable for study with the Levy/Blundell system? Can you please describe these? Is your system basically another version of that system, or are there some nuances here that are different? Either way, this manuscript represents a valuable contribution. However, I wish this were clearer.
c. I was very confused while reading the methods because I am familiar with the Levy/Blundell system and I made some assumptions about the system in this manuscript that were untrue. This was partly because the introduction reads as though the system proposed here improves throughput over Levy/Blundell. My confusion was about the hundreds of barcodes used for the experimental evolution studies. I assumed each of these barcoded strains was in fact a library, containing hundreds of thousands of unique barcodes. But this is not the case, and this system, as depicted, does not seem to improve throughput relative to Levy Blundell. I think the introduction needs to clarify what this systems does and does not do and how it is similar/dissimilar to Levy/Blundell’s.
d. Pooling samples with different barcodes before DNA extraction is a clever idea! This is the kind of thing that might appear earlier in the paper, in some kind of comparison to previous systems. Alternately this could appear in the discussion. Performing the evolutions in 96-well plates is also different from Levy/Blundell and gives the opportunity to study many environments. This might help avoid batch effects, which were very high in Venkatarm et al.
Minor comments:
1. The authors mention Blundell and colleagues in the introduction, but I believe Levy is the co-first author on at least one of the cited papers, maybe both. The text should probably read Blundell, Levy and colleagues.
2. I was mildly confused when reading the methods section referring to the proof of concept experiments because it seemed to me that all of the strains should have the same fitness. Indeed, this was the null expectation. The authors should say so in this section of the methods, perhaps more than once.
3. I love the power analysis! It is exactly the kind of thing I’ve been missing from previous studies.
4. I’m confused about the orange lines in figure 3. It is a little strange that this figure appears before the experimental evolutions are discussed in the results section. Perhaps give a heads up about why the barcodes have different fitness (I assume because they receive beneficial mutations) and why power is weaker than in fitness measurements?
5. Label the control group in figure 5.
Lines 62 – In this paragraph, it might be helpful to address what is actually being measured during these fitness assays (what is fitness in a microbial context and in the traditional fitness assay context?)
Line 69 – correct references to proper format
Line 225-228 – Am I correct that for each evolved population, you pooled the time point 0 and competed it against an unevolved reference strain, and then pooled timepoint 250 and competed the pool against the unevolved reference strain? The wording here is a little confusing.
Figure 1 – could panels be included that illustrates the concept of the barcoding and pooled fitness assays? I think these aspects of the experimental design are more important to illustrate visually than a sampling regime.
There are many papers from the Lang and Desai labs that might be helpful to address sensitivity of fitness measurements, ploidy. For example:
Lang GI, Botstein D, and Desai MM. 2011. Genetic variation and the fate of beneficial mutations in asexual populations. Genetics. Jul;188(3):647-61.
Marad DM, Buskirk SW, Lang GI. 2018. Altered access to beneficial mutations slows adaptation and biases fixed mutations in diploids. Nature Ecology & Evolution. May;2(5):882-889.
Fisher KJ, Buskirk SW, Vignogna RC, Marad DM, Lang GI. 2018. Adaptive genome duplication affects patterns of molecular evolution in Saccharomyces cerevisiae. PLoS Genetics. May 25;14(5):e1007396.
McDonald, M., Rice, D. & Desai, M. Sex speeds adaptation by altering the dynamics of molecular evolution. Nature 531, 233–236 (2016).
A particular comment on the correspondence of barcoding and fluorescence based fitness measurements from Levy et al. 2015 would be helpful
What was the actual representation of the different barcoded strains in the timepoint 0 of the pooled fitness assays (e.g., how easy is it to ensure equal representation across all barcoded strains)? How does the initial abundance influence relative abundance over the course of the assay?
Can the authors address their rationale for including an unevolved reference strain in the competition vs. just measuring relative fitness of the pool by not including a reference strain and just looking at barcode abundance itself?
Line 315 – how do you know there were 20 generations?
Can the authors address their rationale for the paired barcoding in the experimental evolutions? What kind of considerations should be made for how many barcoded strains go in each population to track evolutionary dynamics?
There is some interesting data presented here about how bottlenecking and different environments influence evolutionary dynamics, but this section is very brief and all the data are buried in a bunch of tables and figures in the supplemental. I would suggest a Figure that includes some of this data in the main text. I would also suggest clarifying the biological relevance of the different measurements that were obtained.
Barcoding has been used to track fitness and evolutionary dynamics for many years now, and there are a variety of different systems in use. There is a tradeoff with the proposed system (or any barcoding system), the upfront labor to generate barcoded strains (e.g., transforming and verifying presence of barcodes, validation of neutral fitness) vs. the downstream payoff of pooled fitness assays. This compares to other commonly used systems, like competition against a fluorescently tagged ancestor strain, which requires no upfront labor, but fitness assays are more laborious. Ultimately, it comes down to what questions one is interested in asking. For example, if you were interested in different strain backgrounds, barcoding approaches are challenging. However, the ability to request the strains and use the scripts provided here to analyze data is a big benefit. I can particularly imagine the use of this system in undergraduate courses being very impactful.
Addressing the sensitivity of fitness assays for experimental evolution is an important problem in the field. The authors approach these issues very thoughtfully and methodically. It appears that the sensitivity of barcoded fitness assays presented here is similar to the sensitivity of fluorescent based approaches. In my opinion, the greatest utility of this approach comes from being able to pool strains for fitness assays. This could be particularly helpful for researchers interested in tracking fitness over the course of evolution experiments (e.g., questions about evolvability), which is very difficult to assay with current protocols.
I think my struggle comes down to the framing of the manuscript. The framing is very technical and methods based, but I think it could benefit from highlighting the results and how they are innovative in a more biological context.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.