Bayesian estimation of speciation and extinction from incomplete fossil occurrence data

Area of Science:

• Macroevolutionary biology
• Paleontology
• Computational phylogenetics

Background:

• Species diversity dynamics are governed by speciation and extinction rates.
• Estimating these rates from the fossil record is challenging due to data incompleteness and preservation biases.
• Accurate inference of diversification dynamics requires unbiased estimates of lineage speciation and extinction times.

Purpose of the Study:

• To develop a novel probabilistic framework for jointly estimating species-specific speciation and extinction times.
• To infer time-varying rates of the birth-death process using fossil occurrence data.
• To explicitly model preservation and sampling probabilities to correct for observational biases.

Main Methods:

• A Bayesian algorithm is implemented to jointly estimate speciation and extinction times and rates.
• The model incorporates time-varying speciation and extinction rates, independent of each other.
• Preservation and sampling probabilities are explicitly modeled to infer true lineage lifespans.

Main Results:

• The new framework accurately estimates diversification parameters from simulated data, demonstrating robustness to assumption violations and data incompleteness.
• Application to Rhinocerotidae reveals complex, independent temporal shifts in speciation and extinction rates, explaining the group's rise and fall.
• Estimated birth-death process parameters are comparable to those derived from molecular phylogenies.

Conclusions:

• The developed probabilistic framework provides a robust method for inferring macroevolutionary dynamics from fossil data.
• The model successfully captures complex diversification histories, including rate shifts, and accounts for preservation biases.
• This approach facilitates the integration of fossil and phylogenetic data for a more comprehensive understanding of macroevolutionary processes.