search

tskit-dev / msprime-1.0-paper

Publication describing msprime 1.0
4 stars 20 forks source link

Responses to reviewer 2 & 3 #222

Closed jeromekelleher closed 3 years ago

jeromekelleher commented 3 years ago

These are the main "push-back" responses, so good to get some eyes on this.

It's hard to know how to respond to reviewer 3. At the moment we're saying "no" to everything (to the degree that there's any actionable points in there) but maybe we should meet them in the middle a bit with a sentence about how the neutral theory is an approximation, with some references into the literature?

cc @sgravel @andrewkern @petrelharp @JereKoskela

JereKoskela commented 3 years ago

Thanks @jeromekelleher, this looks very measured to me!

My knowledge of the literature on the applicability of neutral vs non-neutral theory is lacking, but one somewhat-relevant, positive point we could make is that multiple merger coalescents can also arise as models of evolution under positive selection. Their use as non-neutral models is not exactly widespread, but there is at least Neher & Hallatschek (2013) and Hoscheit & Pybus (2019).

sgravel commented 3 years ago

I am at a meeting all day, but a quick take: The main concern of Reviewer 3 is their judgement of impact and fit for the journal. This is to a large extend an editorial call, and based on editorial response I am not too concerned that the editor will hold the ms back for such a judgement call. But it is still worth doing our best effort at addressing the concerns. It is unclear whether the reviewer was reviewing according to the "Methods, technology, and resources" section scope:

"GENETICS welcomes manuscripts that describe genetic or analytic methods or resources that are likely to have broad impact. They can be full length research articles or Communications. The method or resource needs to be novel, or be a significant advance in an existing method or resource. It should be of considerable interest to a wide range of geneticists or of extraordinary interest to a smaller group of geneticists. The method, resource or technology should enable experiments that will allow investigators to address significant biological questions and should be described well enough so others can implement the method. The necessary reagents or resources need to be available upon request."

So it might be helpful to quote from this in the reply and emphasize why we think our work fits in there, rather than starting with emphasizing disagreement and get into a judgement argument about impact.

The reviewer seems to care a lot about inference methods, so highlighting the documented usefulness of msprime to test the robustness of inference methods and identify loci of interest may be helpful?

In general, discussing limitations of neutrality would also not be a bad idea. We could also talk a bit more in the reply about relationship with forward simulators and their limitations, and how there has been a lot of cross-fertilization between the two.

jeromekelleher commented 3 years ago

Going to merge this and make another pass ASAP.