Review

"Differential immune gene expression associated with contemporary range expansion of two invasive rodents in Senegal"

Peer Community in Ecology

November 2018

This ms by Charbonnel and colleagues aims to test predictions associated with the EICA hypothesis (« evolution of increased competittive ability ») in two invasive species. They use transcriptomes from spleen of individuals for populations with different history of invasion (recent versus ancient) and four replicate populations within each invasion history type to find if certain biological functions are differentially activated between recent and ancient populations.

I think that this ms is interesting as the RNA-seq data is used to test an evolutionary model. It is also interesting that the dataset seems to support the opposite of the predicted pattern. The ms is also well written and most analyses are useful and appropriate. However, I have comments on the manuscript that woul dneed to be adressed to streamline it and make it clearer to more generalist readers.

1-The EICA hypothesis is central to the ms but its predictions are not spelled out clearly for non specialists.

The authors should explain in greater details what is proposed by the EICA hypothesis. This could be done on page 3 in the introduction. Specifically, it is not clear what increased competitive ability means, since “success” (or fitness) is not measured in the individuals studied. On page 5, the authors state: “We investigated two alternative hypotheses. On one hand, we expected an overall higher immune gene expression of rodent populations in recently invaded sites, as a response to novel parasite pressures encountered 31 »

This is clear « On the other hand, we expected trade-offs between immune pathways in recently invaded sites under the EICA-refined hypothesis. «

Trade-offs between immune pathways and …what? With other traits (which ones?) . If it is a trade off with other traits such as life history traits that is expected, shouldn’t these life history traits be measured also? Or were the authors expecting to find that trade offs are visible at the gene expression levels, as in Aubin-Horth et al. 2005 (Proc Roy Soc B) who did this in an unrelated study and another tissue in fish? This is especially important for the EICA-refined hypothesis

Or is it « between immune pathways » as in « different immune strategies? « if yes, explain what those different immune pathways are for non specialists. If you expect that different immune strategies have different costs, make a table describing each one with predictions of what is expected

This topic comes back on page 12 in the discussion.

“we did not find any evidence of immunity or particular immune pathways being dampened at the expense of other life history traits or of less energetically costly immune defences. «

Which other life history traits did the authors measure? Not clear.

I understand that the authors found the opposite, with higher immunity expression instead of lower, but still, how could they validate their original prediction with the data they had collected?

On page 10, the authors talk about their predictions “Contrary to what was expected, along the mouse invasion route, all immune-related genes detected were over-expressed in recently invaded sites, and among them, inflammation and complement pathways were over-represented. «

This should be clearer in the introduction

Minor comments INTRODUCTION p.3 “From an eco-evolutionary perspective, invasion success may rely on pre-adaptation within the original range 5,6 or on the rapid evolution of phenotypic traits that would be advantageous in newly colonized areas 5,7. Some supports for this latter process come from the identification of phenotypic variation along invasion gradients, «

I think that phenotypic gradients could be found even if it is based on standing genetic variation, if the alleles are at low frequency in the original population, such that it is difficult to sample the resulting phenotype and / or the alleles that result in the new phenotype are rarely found in the same individuals, but the smaller effective population size of the invading front could result in the « encounter » of these alleles by chance.

RESULTS Even if this information is given in table 1, it would be best to orient the reader by starting the paragraph with 1-tissue studied (spleen) 2-that there are two invasion history (it is really well explained on page 15, maybe pu tat beginning of results?) 3-that there are 4 sites within each invasion history 4-that POOLS of individuals are used. This is important when we try to understand the analysis strategy presented later with the 4vs4 and 8 vs8

The analysis is very complicated with the different ways and different levels of stringency. Could it be possible to only present one?

p.9 the authors use the DE genes in mouse to study protein protein interactions. They focus on the genes related to immunity (using GO terms) and find that they interact within a cluster. Is that a trivial finding / unsurprising? Are there examples of proteins related to the same biological function that have no interaction? What does it tell us more than what we already knew?

TABLE

Table 1 Add a colum on the left with « invasion time » instead of using superscript for each population name

FIGURES FIGURE 1 can be supplementary

Figure 2 should be removed, it is not of enough quality for a public document

Figure 3 I really liked how this figure showed us that even tough there are trends at the average level between invasion histories, there are specific populations that have their own expression profiles for these genes. That is very interesting and will probably warrant more attention in the future.

DISCUSSION p.10 « our results suggested that variations of immune phenotypes were a less important strategy for R. rattus invasion success «

I don’t see where in the ms is invasion success quantified? Could it be that rats are actually not as successful than if they had modulated their immune system? Or that they would be even less successful if they did, since most of the immune activity is up-regulation anyway (which suggest that the invasive mouse is fighting new pathogens)?

p.11 « Although mouse and rat populations experienced reduced genetic diversity due to founder events, they may have developed adaptive responses to novel selection pressures through high levels of plasticity «

Are the authors suggesting that there is genetic accomodation? Please explain in more detail and propose what it means for invasive species in general. Also, this study is a correlation study, and we do not have relationship between the gene expression phenotype and fitness, such that wording should be modulated accordingly.

p.12 « The up-regulation of all immune genes found to be differentially expressed in sites recently invaded by the house mouse strongly supported the assumption of an increased overall infection risk in recently invaded sites. « How do we know this? Please add relevant references

In the manuscript entitled: « Differential immune gene expression associated with contemporary range expansion of two invasive rodents in Senegal » the authors assess the gene expression patterns between anciently and recently established populations of two rodent species in Senegal. They hypothesized that invasion success may relies on immune phenotypic traits that would be advantageous in recently invaded sites. The authors indeed showed that the species Mus musculus domesticus showed an over-expression of immune related genes (notably the complement activation pathway), in recently invaded sites compared to anciently invaded sites and likely related to novel parasite pressures encountered in recently invaded sites. Regarding the species Rattus rattus the authors suggest that some stochastic events may be associated with colonization history since no particular pattern of differential gene expression related to immunity were found.

First, I would like to congratulate the authors for this very interesting work. I found the objectives clear and concise notably with a well written introduction. The analyses are various with a good use of the replicates and seem robust with a deep investigation regarding the biological processes involved. I nevertheless think that the manuscript could be improved notably regarding the structure (see comments below).

Major comments:

1) Even if I found the objectives very clear, I think that the manuscript could still explicit some information earlier in order to gain in clarity. Notably, the whole study focuses on one tissue (the spleen) to analyze gene expression patterns and this information is given a bit late (in the discussion section). However, since the spleen is an immune related tissue I think that this could be explicitly mentioned in the introduction and related to the mains hypothesis investigated, since studying other tissue would test other hypothesis (such as using brain to study behavioral related genes as mentioned page 14).

2) Two approaches are used to test for differentially expressed genes (4x4 and 8x8), however, the subsequent functional analysis mainly focus on the genes identified with the 8x8 approach. I think that the 4x4 approach could be better justified regarding the main questions addressed or better linked to the others results. For instance, are some of the differentially expressed genes of the 4x4 approach related to immune processes?

3) Discussion is a bit redundant (the mention of stochastics events involved in the rat invasion history for instance are discussed at the end of the first paragraph, in the second paragraph and in the fifth paragraph) and could beneficiate from a better structure, notably by addressing the question mentioned in the end of the introduction more directly. For instance in the introduction addressing whether the EICA or the EICA refined hypothesis are supported comes as the last question but is discussed quite early in the discussion (before the functional categories involved that is the second question presented in introduction).

Minor comments

4) Page 6, the number of transcriptome libraries is given, but I think that giving at this stage the number of studied sites and replicates would also ease the comprehension of the subsequent analysis.

5) Page 18, it is mentioned that genes with less than 20 reads or 40 reads were discarded. I suppose that it echoes the “10 occurrences” filter mentioned page 7. If so, I suggest to homogenize and explicit this in order to avoid confusion.

6) Page 6, the authors mentioned that they used a PCA on standardized read counts. I may be wrong, but I think that a multidimensional scaling analysis (MDS) would fit better over dispersed count data (see Bankers et al. 2017, fig. 3 for some example on transcriptomic data).

Bankers, L. et al. 2017. Genomic evidence for population-specific responses to co-evolving parasites in a New Zealand freshwater snail. - Mol. Ecol. 26: 3663–3675.

7) Fig. S1, the PCA is difficult to read, take into consideration to increase front size for instance.

8) Fig. S3 and S4, check carefully the supplementary there is some mismatch between the figure caption for Fig. S3 and S4 and the actual figures displayed.

Typography comments:

• A coma just be added just after “e.g.” and “i.e.”.

• Page 13, remove the double “x” in exhibit.