Bayesian Inference of Evolutionary Histories under Time-Dependent Substitution Rates

Area of Science:

• Evolutionary biology
• Phylogenetics
• Molecular evolution

Background:

• Estimating phylogenetic time scales is challenging due to complex evolutionary models and inference procedures.
• Molecular clock dating reveals that evolutionary rates can vary depending on the time frame of measurement, especially for rapidly evolving viruses.
• This rate variation contrasts with the long-term conservation observed in viruses and endogenous viral elements.

Purpose of the Study:

• To develop a formal and flexible Bayesian statistical inference approach to accommodate time-dependent evolutionary rates.
• To evaluate the novel molecular clock model using diverse biological data, including viral and ancient DNA sequences.
• To compare the performance of the new model against existing molecular clock models.

Main Methods:

• Development of a novel Bayesian statistical inference framework for molecular clock dating.
• Incorporation of time-dependent rate variation within the evolutionary models.
• Application and comparison of the model to foamy virus, lentivirus, Ebolavirus, and woolly mammoth ancient DNA datasets.

Main Results:

• A strong time-dependent evolutionary rate effect was estimated for foamy and lentiviruses, with rates varying over four orders of magnitude.
• The time-dependent effect was weaker but still significant in woolly mammoth ancient DNA, impacting node age estimates.
• Purifying selection at the codon level did not explain the observed time-dependent rates in viruses, though selection influenced Ebolavirus divergence times.

Conclusions:

• The developed Bayesian approach effectively models time-dependent evolutionary rates, improving phylogenetic time scale estimations.
• Time-dependent evolutionary rates are a significant factor across diverse taxa, from viruses to ancient mammals.
• Future work will focus on integrating more complex evolutionary processes to enhance the model's applicability.