Response of phylogenetic diversity to human impacts in terrestrial ecosystems

All code is made available for reproducing the analyses as described in [insert paper URL here]. This is code only, for full completed dataset click here (not yet available).

Pipeline

The analysis comes in four stages:

Each stage is run once with the exception of B1_parse.R, B2_compare.R and B3_metrics.R which are each run twice for comparing the results from unconstrained and constrained pG-lt trees and testing whether there is a difference between total and age branch length normalisation.

Running

Make sure all relevant packages are installed:

Rscript x_install_deps.R

Additionally, you will need a compiled version of pathD8 in your working directory.

To run entire automated pipeline (UNIX OS):

sh x_run.sh

pG-lt step can only be run separately. It was originally run on Imperial's HPC. Make sure you have 0_pglt/ before running the calculate step. C1_analysis.R is an interactive script and it not part of the pipeline. Note, setup will take a long time to run as they interact with the GNR.

Stage details

Setup

This is run with the x_setup.sh script. You can run pG-lt after it is run.

• A1_presolve.R: read in published trees and PREDICTS data, search names against GNR.
• A2_parse.R: read in published trees and rate-smooth
• A3_pgltsetup.R: identify suitable PREDICTS studies, output into pG-lt friendly format. You can run pG-lt with the resulting folder.
• A4_map.R: use published phylogenies in 0_data to generate study-level phylogenies

Calculate

This is run with the x_calculate.R script.

• B1_parse.R: read in pG-lt and mapped phylogenies, check and rate-smooth
• B2_compare.R: compare mapped and pG-lt phylogenies
• B3_metrics.R: generate PD, PSV and PSE values per site using both pG-lt and mapped trees.
• B4_plots.R: community plot PREDICTS data onto pG-lt consensus phylogenies

Directory (after complete run)

-- 0_data/
    -- PREDICTS-DATA/
        -- [site-level PREDICTS data]
    -- raw_trees/
        -- [downloaded published trees]
-- 0_pglt/
-- A1_preresolve/
-- A2_parse/
-- A3_pgltsetup/
-- A4_map/
-- B1_parse/
-- B2_compare/
-- B3_metrics/
-- B4_plots/
-- C1_analysis/
-- stages/
    -- [all R stages files]
-- tools/
    -- [all custom R functions]
-- sanity_checks/
    -- [demonstrations of code accuracy]
-- other/
    -- [misc non-essential items]

Data

After cloning this repo to your machine, you can download all missing data files with this link (not yet available).

Uncompress files:

sh x_uncompress.sh

And compress again:

sh x_compress.sh

Authors

Bennett D.J., De Palma A., Pearse W.D. and Purvis A.