Population structures can be crucial determinants of evolutionary processes. For the Moran process on graphs certain structures suppress selective pressure, while others amplify it (Lieberman et al. 2005 Nature 433 312-316). Evolutionary amplifiers suppress random drift and enhance selection. Recently, some results for the most powerful known evolutionary amplifier, the superstar, have been invalidated by a counter example (Díaz et al. 2013 Proc. R. Soc. A 469 20130193). Here we correct the original proof and derive improved upper and lower bounds, which indicate that the fixation probability remains close to 1 − 1/(r 4 H) for population size N → ∞ and structural parameter H 1. This correction resolves the differences between the two aforementioned papers. We also confirm that in the limit N, H → ∞ superstars remain capable of eliminating random drift and hence of providing arbitrarily strong selective advantages to any beneficial mutation. In addition, we investigate the robustness of amplification in superstars and find that it appears to be a fragile phenomenon with respect to changes in the selection or mutation processes.

As large-scale collaborative, cross-cultural ethnographic research becomes easier and easier to realize, certain ethnographic methods and analyses should be correspondingly more available, inviting, and accommodating. We have therefore created AnthroTools, a package for the free, opensource language R, with a variety of tools and functions suitable for both multi-factor free-list analysis and Bayesian cultural consensus modeling. Free-list data elicitation is a simple technique for ethnographic research. However, especially for cross-cultural free-list data, background preparation is considerable and often requires specific software. In addition, although current cultural consensus analysis tools offer very sophisticated analyses, they also either require specialized software or have computationally taxing methods. AnthroTools expedites these techniques, rapidly performs diagnostics, and prepares data for further analysis. In this article, we briefly discuss what this package offers cross-cultural researchers and provide basic examples of some of its functions.

The rise of the World Airline Network over the past century has led to sharp changes in our notions of "distance" and "closeness"-in terms of both trade and travel, but also (less desirably) with respect to the spread of disease. When novel pathogens are discovered, countries, cities, and hospitals are caught trying to predict how much time they have to prepare. In this paper, by considering the early stages of epidemic spread as a simple branching process, we derive the full probability distribution of arrival times. We are able to rederive a number of past arrival time results (in suitable limits) and demonstrate the robustness of our approach, both to parameter values far outside the traditionally considered regime and to errors in the parameter values used. The branching process approach provides some theoretical justification to the "effective distance" introduced by Brockmann and Helbing [Science 342, 1337 (2013)]; however, we also observe that when compared to real-world data, the predictive power of all methods in this class is significantly lower than has been previously reported.

No abstract

We model the extent to which age targeted quarantine can be used to reduce ICU admissions caused by novel coronavirus COVID-19. Using demographic data from New Zealand, we demonstrate that lowering the age threshold for quarantine to 50 years of age reduces ICU admissions drastically, and show that for sufficiently strict isolation protocols, isolating one third of the countries population for a total of 6 months is sufficient to avoid overwhelming ICU capacity throughout the entire course of the epidemic. Similar results are expected to hold for other countries, though some minor adaption will be required based on local age demographics and hospital facilities.