tRNA functional signatures classify plastids as late-branching cyanobacteria

Area of Science:

• Evolutionary Biology
• Genomics
• Biotechnology

Background:

• Eukaryotes acquired oxygenic photosynthesis via endosymbiosis with a cyanobacterial ancestor of plastids.
• The precise phylogenetic position of plastids within Cyanobacteria remains debated, with conflicting results from phylogenomic studies.
• Existing evolutionary models may not adequately fit complex phylogenomic data, leading to uncertainty.

Purpose of the Study:

• To address the controversy surrounding the phylogenetic root of plastids within Cyanobacteria.
• To develop a robust method for accurate phylogenetic classification of cyanobacterial and plastid genomes.
• To investigate biases in phylogenomic analyses stemming from data heterogeneities.

Main Methods:

• Applied Posterior Predictive Analysis to evaluate the fit of evolutionary models to cyanobacterial and plastid phylogenomic datasets.
• Developed CYANO-MLP, a machine learning algorithm utilizing tRNA gene features for genome classification.
• Tested CYANO-MLP's accuracy and robustness against data biases like compositional heterogeneity and unbalanced sampling.

Main Results:

• Identified significant model-data misfit in existing phylogenomic analyses due to heterogeneities in substitution processes and lineage compositions.
• CYANO-MLP accurately and robustly phyloclassifies cyanobacterial genomes, including plastid genomes.
• Plastid genomes were consistently classified into a late-branching cyanobacterial clade, aligning with metabolic and gene transfer data.

Conclusions:

• Phylogenomic data from cyanobacteria and plastids require sophisticated modeling to overcome site-process and compositional heterogeneities.
• CYANO-MLP offers a robust and accurate phyloclassification strategy, overcoming limitations of traditional methods and amino acid recoding.
• The study strongly supports a late-branching origin for plastids within Cyanobacteria, consistent with multiple lines of evidence.