Feature Frequency Profile (FFP) Method: A Language Model for Whole-Genome (“Organism”) Phylogeny
摘要
For almost half a century, molecular phylogenetic methods (based on genomic sequence variations among a small number of genes/proteins that are highly homologous among the study population) have dominated in inferring the evolution of diverse organisms despite many important intrinsic limitations. Strictly speaking, these molecular trees represent the inferred evolution of those molecules selected, neither of organisms, nor of whole genomes (as simplified yet quantifiable proxy for the organism) selected. This is primarily because of two reasons: (a) the technology of economic sequencing and assembling the whole genome of an organism did not emerge until recently, especially for eukaryotic whole genomes, and (b) computational methods of comparing whole-genome sequences efficiently did not emerge until recently due to the sparsity of available whole-genome sequences. An example of a whole-genome Tree of Life is presented.