A phylogenetic Kalman filter for ancestral trait reconstruction using molecular data

Area of Science:

• Evolutionary biology
• Genomics
• Computational biology

Background:

• Correlating life history/ecological traits with genomic features aids evolutionary history reconstruction.
• Current methods using linear regression neglect phylogenetic inertia.
• Comparative regression offers a more robust framework for ancestral trait reconstruction.

Purpose of the Study:

• To develop a more efficient and accurate Bayesian method for comparative regression in phylogenetics.
• To address limitations of existing generalized least-square approaches.
• To reconstruct the evolutionary history of traits, such as optimal growth temperature in Archaea.

Main Methods:

• Introduced a Bayesian sampler as an alternative to generalized least-square methods.
• Formally equated ancestral trait reconstruction with a phylogenetic Kalman filter problem.
• Developed and implemented forward-backward recursions within a Markov chain Monte Carlo sampler.

Main Results:

• The novel comparative regression method yields more accurate ancestral reconstructions.
• The method provides a more faithful representation of uncertainty compared to simple linear regression.
• Application to Archaea's optimal growth temperature suggests a hyperthermophilic ancestor, consistent with prior findings.

Conclusions:

• The Bayesian approach offers a significant algorithmic improvement for phylogenetic comparative regression.
• Accurate ancestral state reconstruction is crucial for understanding trait evolution.
• The findings support the hypothesis of a hyperthermophilic origin for the Archaea kingdom.