Complicated history of gene duplication and loss brings challenge to molecular phylogenetic inference, especially in deep phylogenies. However, phylogenomic approaches, such as gene tree parsimony (GTP), show advantage over some other approaches in its ability to use gene families with duplications. GTP searches the ‘optimal’ species tree by minimizing the total cost of biological events such as duplications, but accuracy of GTP and phylogenetic signal in the context of different gene families with distinct histories of duplication and loss are unclear.    

To evaluate how different evolutionary properties of different gene families can impact on species tree inference, SHI Tao, from Plants Reproductive Biology Research Group, Wuhan Botanical Garden, analyzed 3900 gene families from seven angiosperms encompassing a wide range of gene content, lineage-specific expansions and contractions.  

Study revealed that the gene content and total duplication number in a gene family strongly influenced species tree inference accuracy, with the highest accuracy achieved at either very low or very high gene content (or duplication number) and lowest accuracy centered in intermediate gene content (or duplication number), as the relationship can fit a binomial regression. For gene families of similar level of average gene content, those with relatively higher lineage-specific expansion or duplication rates tend to show lower accuracy.  

Additional correlation tests supported that high accuracy for those gene families with large gene content may rely on abundant ancestral copies to provide many subtrees to resolve conflicts, whereas high accuracy for single or low copy gene families were just subject to sequence substitution per se. Very low accuracy reached by gene families of intermediate gene content or duplication number could be due to insufficient subtrees to resolve the conflicts from loss of alternative copies.  

As these evolutionary properties could significantly influence species tree accuracy, researchers would discusse the potential weighting of the duplication cost by evolutionary properties of gene families in future GTP analyses. 

Results were published in Molecular Phylogenetics and Evolution entitled “Impact of gene family evolutionary histories on phylogenetic species tree inference by gene tree parsimony”.  

 Properties such as gene family size and variation of gene content among taxa strongly influencethe species tree inference accuracy.(Image by SHI Tao)