iBet uBet web content aggregator. Adding the entire web to your favor.
iBet uBet web content aggregator. Adding the entire web to your favor.



Link to original content: http://pubmed.ncbi.nlm.nih.gov/39047110/
Trait-mediated speciation and human-driven extinctions in proboscideans revealed by unsupervised Bayesian neural networks - PubMed Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2024 Jul 26;10(30):eadl2643.
doi: 10.1126/sciadv.adl2643. Epub 2024 Jul 24.

Trait-mediated speciation and human-driven extinctions in proboscideans revealed by unsupervised Bayesian neural networks

Torsten Hauffe  1 , Juan L Cantalapiedra  2   3   4 , Daniele Silvestro  1   5
Affiliations

Trait-mediated speciation and human-driven extinctions in proboscideans revealed by unsupervised Bayesian neural networks

Torsten Hauffe et al. Sci Adv. .

Abstract

Species life-history traits, paleoenvironment, and biotic interactions likely influence speciation and extinction rates, affecting species richness over time. Birth-death models inferring the impact of these factors typically assume monotonic relationships between single predictors and rates, limiting our ability to assess more complex effects and their relative importance and interaction. We introduce a Bayesian birth-death model using unsupervised neural networks to explore multifactorial and nonlinear effects on speciation and extinction rates using fossil data. It infers lineage- and time-specific rates and disentangles predictor effects and importance through explainable artificial intelligence techniques. Analysis of the proboscidean fossil record revealed speciation rates shaped by dietary flexibility and biogeographic events. The emergence of modern humans escalated extinction rates, causing recent diversity decline, while regional climate had a lesser impact. Our model paves the way for an improved understanding of the intricate dynamics shaping clade diversification.

PubMed Disclaimer

Figures

Fig. 1.
Fig. 1.. Comparing simulated and inferred lineage-time–specific rates.
A single simulation under four diversification scenarios (settings listed in fig. S1) is displayed to exemplify the accuracy and coverage of rates inferred by the BDNN model. Dashed lines indicate simulated speciation and extinction rates. Dots display the means, and vertical solid lines display the 95% credible interval (CI) of the inferred rates. Accuracy was quantified by the median absolute percentage error between the simulated and inferred rates, and the coverage gives the share of lineages whose CI includes the simulated rate. The true simulated speciation rate was obtained from the ancestral lineage from which the lineage descended, from which the BDNN model infers the rate. All rates are given in units of events lineage−1 Myr−1.
Fig. 2.
Fig. 2.. Power to find evidence of rate variation among species and over time.
We quantified the coefficient of variation in inferred lineage-specific speciation and extinction rates, respectively. Across 100 thresholds, we calculated the proportion of correctly evidenced rate variation when simulated rates varied (i.e., mean sensitivity across scenarios 2 to 9) and erroneously inferred rate variation when, in fact, diversification under equal and constant rates was simulated (i.e., specificity for scenario 1). Dashed vertical lines display the thresholds for speciation and extinction above which a constant-rate model was rejected with 95% specificity.
Fig. 3.
Fig. 3.. Inferred effects of selected predictors on speciation and extinction rates.
PD plots showing the inferred effect of traits and time-dependent variables that actually influenced simulated speciation and extinction rates while marginalizing over predictors without a simulated impact. Dashed lines visualize the simulated effect, transparent lines show the 100 simulated replicates, and the thick solid line represents the average across them after locally estimated scatterplot smoothing (loess). Height of the barplots displays the proportion of simulations that exceeded our threshold to detect rate variation (i.e., rejecting the constant-rate model). The share of dark gray indicates how often the correct traits or time-dependent variables were correctly identified (when included among the predictors), and the white proportion shows incorrectly identified predictors. Light gray displays cases where the predictive time series was not included and instead time or phylogenetic relatedness was detected (scenarios 3 and 8) or only one of the two predictors was found (scenarios 4 and 7). Unit of rates are events lineage−1 Myr−1.
Fig. 4.
Fig. 4.. Analysis of the proboscidean fossil record using the BDNN model.
(A) The inferred species-specific speciation and extinction rates show substantial heterogeneity, as showcased by the nine species displayed here. (B and D) Changes in speciation rates are mostly driven by time, geography, and an ecomorphological trait related to diet, with the highest rates found in the Oligocene and Miocene species with generalist diet, and with a consistent increase in island versus continental species. (C) Extinction rates were primarily modulated by the spatiotemporal overlap with early and modern humans, with additional effects of an ecomorphological trait related to mandible and tusk shapes, and (D) with higher extinction in island species. Unit of rates are events lineage−1 Myr−1. Silhouettes are made available on PhyloPic.
See this image and copyright information in PMC

Similar articles

See all similar articles

References

    1. Blanco F., Calatayud J., Martín-Perea D. M., Domingo M. S., Menéndez I., Müller J., Fernández M. H., Cantalapiedra J. L., Punctuated ecological equilibrium in mammal communities over evolutionary time scales. Science 372, 300–303 (2021). - PubMed
    1. Fraser D., Villaseñor A., Tóth A. B., Balk M. A., Eronen J. T., Andrew Barr W., Behrensmeyer A. K., Davis M., Du A., Tyler Faith J., Graves G. R., Gotelli N. J., Jukar A. M., Looy C. V., McGill B. J., Miller J. H., Pineda-Munoz S., Potts R., Shupinski A. B., Soul L. C., Kathleen Lyons S., Late quaternary biotic homogenization of North American mammalian faunas. Nat. Commun. 13, 3940 (2022). - PMC - PubMed
    1. Rolland J., Condamine F. L., Beeravolu C. R., Jiguet F., Morlon H., Dispersal is a major driver of the latitudinal diversity gradient of Carnivora. Glob. Ecol. Biogeogr. 24, 1059–1071 (2015).
    1. Rabosky D. L., Chang J., Title P. O., Cowman P. F., Sallan L., Friedman M., Kaschner K., Garilao C., Near T. J., Coll M., Alfaro M. E., An inverse latitudinal gradient in speciation rate for marine fishes. Nature 559, 392–395 (2018). - PubMed
    1. Fenton I. S., Aze T., Farnsworth A., Valdes P., Saupe E. E., Origination of the modern-style diversity gradient 15 million years ago. Nature 614, 708–712 (2023). - PubMed

LinkOut - more resources