02 - InvertR - Assignment 1 - Dispersal - Example in R

Overview

Alex Smith uses a South African velvet-worm dataset to demonstrate an R workflow for inferring dispersal from DNA sequences and geographic locations, while explaining how the same methods apply to the assignment’s deep-sea hydrothermal-vent scenario. The workflow downloads GenBank sequences from accessions, aligns them, builds and annotates a maximum-likelihood phylogeny, tests isolation by distance with a Mantel test, and produces maps and plots for a three-minute assignment presentation.

Key takeaways

The South African velvet-worm workflow is a methodological stand-in for the hydrothermal-vent assignment: the habitat changes, but the sequence, geography, and dispersal analysis steps transfer.
GenBank accessions connect individual DNA sequences to collection locations, allowing genetic relationships to be interpreted alongside geographic separation.
Sequence alignment must precede phylogenetic and distance analysis because unaligned sequences may compare nonhomologous positions.
A Mantel test is appropriate for testing association between pairwise genetic and geographic distance matrices because those pairwise observations are not independent.
The example combines a significant positive isolation-by-distance relationship with deep within-species tree branches to infer restricted dispersal among sites.
The assignment’s final deliverable is an annotated, three-minute video presentation—not simply the generated PDF—and it should relate the figures to hydrothermal-vent vulnerability and mining pressure.

Chapters

0:00 Why R, Reproducible Science, and the Assignment Setup

R is a free statistical and programming language widely used in science, where publishing code supports reproducibility.
Alex Smith recommends installing the current R version from CRAN before installing RStudio, which provides the shared coding environment.
Download the assignment’s GitHub repository as a ZIP and extract its R script and CSV data into an R working directory.

5:50 Hydrothermal-Vent Assignment and the South African Velvet-Worm Example

The assignment uses genetic evidence to estimate larval dispersal among deep-sea hydrothermal vents and consider implications for isolation, fragmentation, and mining pressure.
The demonstration uses publicly available South African Onychophora data so students can practice methods that transfer across terrestrial and marine habitats.
Published site coordinates and DNA accessions let students reanalyze data associated with research on velvet-worm diversity around Cape Town.

10:50 RStudio’s Four Panes and Step-by-Step Script Workflow

The Source pane displays the R script; lines beginning with # are explanatory comments, while uncommented lines execute code.
The Console shows commands, output, and errors; the Environment tracks imported objects; and the Plots pane displays generated figures.
Smith recommends running code in sequence, one block at a time, and checking whether warnings actually prevent later steps from running.

17:15 Set the Working Directory and Install R Packages

Use get working directory to see where R expects files; change the directory only if the extracted GitHub files are somewhere else.
Clear the R environment before starting to avoid accidentally reusing objects with names that conflict with the example.
Install required packages, including through BiocManager, while online; then call library commands to load them for sequence analysis, mapping, and plotting.

21:00 Import Site Data and Map Velvet-Worm Collection Locations

Read the CSV into R; its columns include GenBank accession, site description, latitude, longitude, and species name.
Convert coordinates from degrees, minutes, and seconds to decimal degrees, then create a site map using sf, Natural Earth data, and ggplot2.
Make both a world map and a zoomed-in map showing five collection sites along the South African coast.

25:00 Retrieve GenBank Sequences, Align Them, and Build a Phylogeny

Use the CSV’s accession numbers to retrieve DNA records from GenBank; the example sequences are mitochondrial cytochrome c oxidase I (COI).
Save sequences in FASTA format and align them before analysis so that comparisons involve homologous regions rather than offset sequence positions.
Build a maximum-likelihood tree, root it at the midpoint because no outgroup is supplied, and interpret genetic distances from horizontal branch lengths.

33:30 Annotate the Phylogeny with Species and Collection Sites

Create a site-by-accession presence–absence matrix, using 1 for a sample present at a site and 0 for absent.
Color tree tips by species and add geographic site information with tree-plotting tools such as ggtree and ggtree spatial functionality.
The annotated tree shows four named taxa and highlights deep genetic variation among samples assigned to one species across multiple sites.

37:35 Measure Isolation by Distance with a Mantel Test

Subset the data to the purple focal species and calculate pairwise geographic distances among its collection sites in kilometers.
Compare geographic and genetic distance matrices with a Mantel test from vegan, which accounts for pairwise observations that are not independent.
For the example species, genetic distance increases with geographic distance, and the Mantel result is significant; an additional regression plot visualizes the relationship.

42:15 Export Three Assignment Figures and Prepare the Presentation

Export a PDF containing the site map, the species- and site-annotated phylogeny, and the isolation-by-distance plot.
The assignment submission is not the PDF: students should annotate the figures and record a separate three-minute video answering four questions.
Use the scheduled Friday lab time to troubleshoot R, interpret the geographic, phylogenetic, and isolation-by-distance results, and connect them to hydrothermal-vent dispersal.

Summary, takeaways, and chapters were generated by AI from the video's transcript and may contain errors. The video belongs to its creator, Alex Smith.