Skip to contents

pfwim: Paleo Food Web Inference Model

pfrim is an R package for inferring predator–prey interactions using trait-based rules and generating hypothetical realised food webs. It implements the Paleo Food-web Inference Model (Shaw 2024) and includes tools for categorical trait matching, numerical predator–prey size rules, and power-law downsampling of realised webs.

Installation

Install from CRAN

You can install the development version of pfwim from GitHub with:

# install.packages("pak")
pak::pak("BecksLab/pfwim")

Examples 

library(pfwim)

# Trait data
data("traits", package = "pfwim")

# Trait combination rules
data("feeding_rules", package = "pfwim")

head(traits)
head(feeding_rules)

Inferring a consumer–resource edgelist

Use infer_edgelist() to infer feasible interactions:

edgelist <- infer_edgelist(
  data = traits,
  cat_combo_list = feeding_rules,
  col_taxon = "species",
  certainty_req = "all",
)

# Show first few interactions
head(edgelist)

Generate hypothetical realised webs

Create a series of hypothetical webs using powerlaw_prey()

webs <- powerlaw_prey(
  el = edgelist,
  n_samp = 5,
  y = 2.5
)

# Inspect first web
webs[[1]]

Each element of webs is a matrix representing a realised food web generated from the inferred interactions

Citation

If you use pfwim in your research, please cite the associated manuscript:

Shaw, Jack O., Alexander M. Dunhill, Andrew P. Beckerman, Jennifer A. Dunne, and Pincelli M. Hull. 2024. A Framework for Reconstructing Ancient Food Webs Using Functional Trait Data. bioRxiv. https://doi.org/10.1101/2024.01.30.578036.