Computational protocol: Substantial variation in the timing of pollen production reduces reproductive synchrony between distant populations of Pinus sylvestris L. in Scotland

Similar protocols

Protocol publication

[…] All the statistical analyses were performed in R version 3.2.3. (R Core Team, ). Data management, analysis, and visualization relied upon the “dplyr” (Wickham & Francois, ) and “ggplot2” (Wickham, ) packages.Due to the time intervals between site visits, it was not always possible to be at each of the sites at precisely the time when the majority of pollen is shed, a period which, in Finland, lasts only around 3 days per tree (Parantainen & Pulkkinen, ). To overcome this, estimates of the differences in timing of development between sites were made using cumulative link models, a type of ordinal logistic regression implemented using the “ordinal” package within R (Christensen, ). The purpose of a cumulative link model is to estimate the cumulative probability that a given observation will fall into one of a series of ordinal categories based on predictor terms provided in model specification. A major advantage of ordinal regression models for this purpose is that they recognize that an ordinal response is bounded at both ends and make no assumption about the spacing between values of the response variable, as would be implied by a linear regression model with a continuous response (Harrell, ). […]

Pipeline specifications

Software tools dplyr, Ggplot2
Application Miscellaneous