Jake Berv

@jakeberv.bsky.social

451 Followers 588 Following

Schmidt AI in Science Postdoctoral Fellow | evolution | systematic biology | natural history | barbeque enthusiast

Supplementary Figure 17. Quantile occupancy and tail-composition diagnostics. (A–B) Quantile occupancy maps showing the number of species per H3 grid cell belonging to the global fast25 (top quartile of lineage rates; A) and slow25 (bottom quartile; B) sets. Each species contributes +1 to every cell it occupies. (C–D) Corresponding compositional maps showing, for each cell, the fraction of the local assemblage comprised of fast25 (C) or slow25 (D) lineages (0–1 scale). Mean lineage rates are positively correlated with the fast25 fraction (E) and negatively correlated with the slow25 fraction (F); dashed vertical lines indicate the global expectation under the quartile definition (0.25). (G–H) Cells in the highest decile of mean lineage rates are enriched for fast25 lineages (G) and depleted of slow25 lineages (H), visualized as violin plots with embedded boxplots; dashed horizontal lines indicate the 0.25 expectation. Some tropical regions, including the Amazon basin, show substantial representation from both fast- and slow-rate tails, whereas others, including parts of central Africa, appear less strongly enriched for either extreme; compositional maps also suggest a transition across Wallacea. Because these summaries are based on year-round resident distributions, such contrasts may reflect the combined effects of climatic history, dispersal barriers, lineage turnover, and sampling structure.
Spatial predictors of phenotypic evolution across global assemblages. a, Predicted grid-cell mean multivariate phenotypic rates (inverse range weighted; Methods), from an aggregated SAR-Lag model including key predictors: temperature seasonality (BioClim4, °C × 100), species richness, and latitude, as well as additional controls for sampling and speciation rates (Methods). Estimated (Fig. 1a and Supplementary Figs. 8 and 9) and predicted mean rates are typically higher at higher latitudes and are well approximated by a simple quadratic relationship with latitude (Supplementary Fig. 13). b, Temperature seasonality (BioClim4) shows spatial patterns similar to a. c, Aggregated SAR-Lag model (in a) residuals projected across space, showing minimal residual structure after accounting for spatial autocorrelation.
Supplementary Figure 18. from Berv et al 2026. Figure 1 is reproduced, with enlarged illustrations that are approximately to scale relative to one another and suitable for printing. Illustrated species names are noted in Supplementary Table 7.
Matt Ferguson, Jurassic Park, They Do Move in Herds
Announcement for a new Biodiversity Informatics Research Group at UNC Charlotte, launching January 2027. Dr. Jake Berv is launching a lab studying biodiversity across space and deep time, integrating phylogenomics, fossil and modern specimens, fieldwork, and comparative methods. A major near-term interest is developing and benchmarking human-guided AI research agents for rigorous, reproducible analyses in ecology and evolution. Prospective students, postdocs, and collaborators in macroevolution, computational biology and AI, natural history, and open science are welcome to get in touch at jacob.berv@charlotte.edu. More information: jakeberv.com. The graphic features McDowell Nature Preserve in Charlotte.