mosuo-ppr

Analysis script for "Offspring sex preferences among patrilineal and matrilineal Mosuo in Southwest China revealed by differences in parity progression" by Siobhán M. Mattison, Bret Beheim, Bridget Chak, Peter Buston R. Soc. open sci. 2016 3 160526; DOI: 10.1098/rsos.160526. Published 14 September 2016

Requirements:

• R (3.3.1 or greater) https://cran.r-project.org/
• rethinking package (v1.59 or greater), http://xcelab.net/rm/software/

Instructions:

In R, set the working directory to that containing this readme file. For example, on a Mac or Linux machine, you might say

    setwd('~/Desktop/mosuo-ppr')

if the folder containing the project is named 'mosuo-ppr' and on your Desktop. You can tell if you are in the right place by typing in dir() and seeing the folders 'code' and 'inputs' and this readme.txt file. The analysis takes as input two files:

'final_regression_data.csv' - this file contains relevant information about our sample of women, including their reproductive histories.

'Mosuo_pop_reg.csv' - this is a pedigree file, so it contains both men and women, indexed by random alphanumeric code 'pid'. This is used as the starting point for the ppr simulation - we grow the population forward 100 years and observe the consequences of living under the fertility regime estimated in our models, and then empirically calculate the honeycomb PPR numbers.

The analysis itself is broken up into independent modules that pass outputs to each other. The whole process runs by typing one command into R,

source('./code/runproject.r')

with the project folder as the working directory. If all goes well, each step of the analysis will execute in sequence, and write the final tables and figures into an 'output' folder, along with a runtime log.

By default the analysis will delete all temporary files and folders, but if you want to see all intermediate steps you can disable this by flipping the save_temp variable in 'project_variables.r' from FALSE to TRUE.

The total time until completion will vary by machine; the slowest we have seen has been about 2 hours.

The project is maintained by Bret Beheim (beheim at gmail dot com) and is hosted at https://github.com/babeheim.