Molecular evolutionary signatures reveal the role of host ecological dynamics in viral disease emergence and spread

Dule-Sylvester et al., 2013. Philosophical Transactions of the Royal Society B 368 – 1614. DOI:10.1098/rstb.2012.0194. Molecular evolutionary signatures reveal the role of host ecological dynamics in viral disease emergence and spread

Below, we give our first impressions of this article. Please comment below, or tweet Will or Lynsey (maybe use #pegejc). Think of this as a journal club discussion group!

Will Pearse

Will Pearse

We covered this paper in a (real, live, in-person) journal club here at the University of Minnesota, so the views below are probably not just mine. So, if you like what’s written below, it’s from the community genetics reading group; if you don’t like what’s written below, it’s all from me.

This study links an epidemiological model of how rabies spreads among raccoons to the structure of the genealogy that the rabies would have given present-day sequence data. I really like the modeling framework this paper presents. Inferring ecological patterns directly from a genealogy is a brave thing to try and do, and while others have done similar work this is probably the most explicit model I’ve seen someone trying to fit. This is also the paper’s greatest weakness: I’m not sure that these models could ever be fit successfully to real data. Taking figure 3 as an example, the authors state that they can detect the influence of long-distance dispersal because the exponential growth model fits their data better; I don’t think we would ever get such neat graphs with real data, and the predictions of their linear and exponential models look too similar (to me) to distinguish between in the presence of experimental noise. Indeed, while the authors use parameters derived from real data, they don’t actually attempt to fit their models to real genetic data; I wonder if they would be able to do so.

Moving past those rather snarky comments, this paper interested me because they’re attempting to model the ecological processes that might produce a particular genealogical (phylogenetic) structure. By looking for what kinds of signals long-distance dispersal leaves in the genome of rabies, they’re able to make useful predictions about what the rabies is doing right now – that’s presumably a lot of help if you’re trying to control an epidemic. I’d never really thought about how perfect a system diseases are for eco-phylogeneticists – they jump from host-to-host, making lineages nice and separate, and they evolve really quickly. Let’s just ignore multiple infections and DNA saturation for a moment, and think about the opportunities for fitting these kinds of complex models. Maybe we can all start linking phylogenetic (whoops – genealogical) structure to explicit models of evolution that incorporate ecology, and in the process help better-understand disease dynamics. As an eco-phylogeneticist, that kind of excites me!

Lynsey McInnes

Lynsey McInnes

First, apologies for the delay to this week’s post – I got caught up in Easter Monday laziness and what follows is largely random thoughts that popped into my mind as I read this paper on the train into work this morning.

I really enjoyed the idea behind this paper. I haven’t read much of the literature around the eco-evolutionary dynamics of virus evolution, but it sounds like crazy fun. I have read A LOT of the literature around models of spatially-explicit diversification and this paper definitely made me want to see more cross-talk between these two research areas (neatly incorporating my new field of statistical phylogeography/population genetics).

(I think) just like Will, I was excited by the possibilities that the authors outline, impressed by their modelling framework, but dubious about some of their outcomes and the likelihood that such a detailed model could often be used for predictive inference. I’d be happy to proven wrong however, and have very little feeling for how much data is really need/exists for such models to be powerful for, e.g. public health decision making. I’m also not convinced by figure 2 – is there not a ton of pseudoreplication going on in there – should there not be only five data points (as in figure 1b). Dare I say it – how about a mixed effects model?

Although the authors did perform sensitivity analyses and spend time discussing the effects of landscape heterogeneity and demographic stochasticity on their ability to infer process, I would have liked to see have seen more exploration of the effect of missing or biased data (for example how noisy can the data be before signal becomes distorted/lost?). I concede I have not checked the supplementary information and this information might be in there…

As a side project, I’ve been thinking about the effects of dispersal on macroevolutionary diversification and it was refreshing in this paper to see local and long distance dispersal so simply made distinct. I think this clarity of distinction is lacking from macroevolutionary analyses (so that when people look for the effects of dispersal on diversification they get conflicting results depending on whether they are looking inside a restricted area or beyond it (to cut a long story short)). Here, the authors have clear hypotheses on the differences expected whether or not the host moves beyond its immediate neighbourhood. One imagines that there isn’t really two distinct categories, but there is certainly more than one. So, hooray.

This comment might have come across as overly negative. I did not mean it to. I really enjoyed reading this paper, it was extremely well-written and thought-provoking (such that someone with no real background in disease dynamics could understand both the rationale and methods). I am going to check out the other articles in the special issue of Phil Trans that this article came from and look forward both to seeing how these types of models develop and hopefully to pilfering some of these ideas across into macroevolutionary diversification (that is similarly affected by processing acting on ecological time-scales (always good to end on a blatant note of self-citation).


About will.pearse
Ecology / evolutionary biologist

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: