Commit a7e859e4 authored by Jerome Waldispuhl's avatar Jerome Waldispuhl
Browse files

tmp ref 2

parent 54ea07f3
......@@ -67,7 +67,7 @@
First of all, we would like to thank the reviewers for their careful review of our manuscript and precise comments. We do appreciate the time invested in reviewing our manuscript and quality of their remarks.\\
However, we believe that misunderstandings of our methodologies and claims may have negatively altered the editorial decision. We take full responsibility for any ambiguity. Although, we would like to stress that our hypothesis and findings are already strikingly inline with the reviewer comments. For this reason, we believe that a clarification of our methodology and claims could motivate a positive re-evaluation of the impact of our work, and the possibility to re-submit a revised version of our manuscript.\\
However, we believe that misunderstandings of our methodologies and claims may have negatively altered the editorial decision. We take full responsibility for any ambiguity. Although, we would like to stress that \textbf{our hypothesis and findings are already strikingly inline with most of the reviewer comments}. For this reason, we believe that a clarification of our methodology and claims could motivate a positive re-evaluation of the impact of our work, and the possibility to re-submit a revised version of our manuscript.\\
In this letter, we clarify these misunderstandings and highlight key arguments that might have been missed in our original manuscript. Before addressing all comments, we remind the scenarios explored in our study:
......@@ -77,19 +77,21 @@ In this letter, we clarify these misunderstandings and highlight key arguments t
\end{enumerate}
Importantly, we did not investigate natural selection scenarios based on the fitness of a population to target(s) phenotype(s). It has been extensively studied in the past \cite{???}, and we justified this choice in the introduction (page ???).\\
We identified four major misunderstandings that likely resulted in low appreciation of our work.\\
We identified four major misunderstandings that likely resulted in a low appreciation of our work.\\
\claimstag There is a misunderstanding of the claims made in this manuscript. In particular, as stated on page~5, we do not advocate for an energy-based natural selection model (i.e. \maternal). Instead, we support the hypothesis that RNA were randomly replicated and that stable complex structures have been spontaneously discovered (See page~4).\\
\experimentstag Comparison and experiments with an uniform sampling model appears to have been (unintentionally) overlooked (See Fig.~6).\\
\experimentstag Results and discussions of experiments conducted with an uniform sampling model appears to have been (unintentionally) overlooked (See Fig.~6).\\
\hypothesistag The principles of the evolutionary scenario supported in this study have not been correctly interpreted.\\
\hypothesistag The principles of the evolutionary scenario supported in this study have not been correctly interpreted. It might results from a misunderstanding of the \rnamutants algorithm \cite{Waldispuhl:2008aa} used in this study.\\
\methodstag The principles and output of \rnamutants \cite{Waldispuhl:2008aa}, which has been used to characterized the energy landscape, are misundertood.\\
We strongly believe that the misunderstandings cited above can be easily clarified in a revised version of the manuscript. Please, find below our detailed answers to the reviewers.
%%% REVIEWER 1
\Reviewer{\#1}
\begin{response}{
......@@ -164,17 +166,24 @@ Just after the Z equation it says a pair (l,w), but the equation is for P(w,s).
Yes. We thank the reviewer for this observation that has been fixed in the revised manuscript.
\end{response}
%%% REVIEWER 2
\Reviewer{\#2}
\begin{response}{
In this work the authors explore the space of RNA sequences of length 50 through random sampling and through an evolutionary algorithm. They characterize some structural properties of the corresponding secondary structures paying attention in particular to the folding energy and to the GC content. This is a computationally intensive work that relies on efficient algorithms previously developed by the senior author of this work and other co-workers. While I do not have any technical objection to this study, I have found have major conceptual issues regarding the context and motivation of the study and the interpretation of the results.}
\methodstag There is a deep misunderstanding on how \rnamutants works. As discussed in Sec.4.2. it is not random but instead a weighted sampling of pairs sequence-stucture where the weight is the energy of the sequence folded in that particular structure, compared to energy of all other pairs sequence-structure such that the sequence is at distance $k$ from the wild type.
\methodstag We believe that a misunderstanding of the purpose and output of \rnamutants~\cite{Waldispuhl:2008aa} led to misinterpretation of the results and impact of our work. First, in order to remove any ambiguity, we emphasize that \rnamutants does not sample mutants from a uniform distribution. Instead it extracts in priority mutants folding into stable structures (See Section~4.2). This enabled us to characterize the distribution of stable structure in the mutational landscape.
\end{response}
\begin{response}{
My main criticism regards the ``distinct region of the sequence landscape enriched with multi-branched structures''. I do not see in which sense is this ``region'' special, and in my opinion the ms does not clarify this point. To begin with, the authors probably know that the abundance of structural motifs in (minimum free energy) RNA secondary structures varies with the length of the sequence. This is why short sequences are rich in hairpins and poor in multi-loops, which start to be frequent and lengths about 50, and become unavoidable for longer RNA sequences.}
\experimentstag We do agree. In fact, the data presented in Figure~6 (uniformly sampled sequences represented in dotted lines) already illustrates this claim.
\end{response}
\begin{response}{
My main criticism regards the ``distinct region of the sequence landscape enriched with multi-branched structures''. I do not see in which sense is this ``region'' special, and in my opinion the ms does not clarify this point. To begin with, the authors probably know that the abundance of structural motifs in (minimum free energy) RNA secondary structures varies with the length of the sequence. This is why short sequences are rich in hairpins and poor in multi-loops, which start to be frequent and lengths about 50, and become unavoidable for longer RNA sequences. The discussion in the ms is confusing at several points because different lengths are mixed to discuss what is typical and what is not. In the absence of energetic considerations, the abundance of any structural motif converges to a Gaussian distribution for sufficiently long sequences (50 is already sufficiently long), and therefore typical shapes exist and can be defined. Check for instance Nebel, M.E., 2002. Combinatorial properties of RNA secondary structures. J. Comp. Biol. 9, 541?573 or Poznanovi?, S., Heitsch, C.E., 2014. Asymptotic distribution of motifs in a stochastic context-free grammar model of RNA folding. J. Math. Biol. 69, 1743?1772. In the light of those results the main question is: in which sense do structures in the ?distinct region? differ from typical structures at those lengths? Can be those differences, if present at all, be ascribed to selection for low-energy folds? How are those results related to the apparent lack of structural selection identified by other authors (their ref. [18]), where most abundant structures seem to be the ones found in non-coding RNA databases? This main criticism takes different forms along the ms and in the light of a number of statements and interpretations, as follows.}
\methodstag We agree with the reviewer that if the sampling was random then those results would hold. But as discussed previously it is not random but instead a weighted sampling.
The discussion in the ms is confusing at several points because different lengths are mixed to discuss what is typical and what is not. In the absence of energetic considerations, the abundance of any structural motif converges to a Gaussian distribution for sufficiently long sequences (50 is already sufficiently long), and therefore typical shapes exist and can be defined. Check for instance Nebel, M.E., 2002. Combinatorial properties of RNA secondary structures. J. Comp. Biol. 9, 541?573 or Poznanovi?, S., Heitsch, C.E., 2014. Asymptotic distribution of motifs in a stochastic context-free grammar model of RNA folding. J. Math. Biol. 69, 1743?1772. In the light of those results the main question is: in which sense do structures in the "distinct region" differ from typical structures at those lengths? Can be those differences, if present at all, be ascribed to selection for low-energy folds? How are those results related to the apparent lack of structural selection identified by other authors (their ref. [18]), where most abundant structures seem to be the ones found in non-coding RNA databases? This main criticism takes different forms along the ms and in the light of a number of statements and interpretations, as follows.}
\methodstag
We agree with the reviewer that if the sampling was random then those results would hold. But as discussed previously it is not random but instead a weighted sampling.
\end{response}
\begin{response}{
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment