Correction: Can quartet analyses combining maximum likelihood estimation and Hennigian logic overcome long branch attraction in phylogenomic sequence data?

There are multiple errors in Fig 1. The authors have provided a corrected version here. The publisher apologizes for the error.

Fig 1. Flowchart of the PhyQuart algorithm.

Simplified flowchart showing a) each of the three possible quartet relationships for a set of 4 sequences (q₁, q₂, q₃), b) the site-pattern classification of observed (N_obs) symmetric (${\xi }_{q_i}$) and asymmetric (${\sigma }_{q_i}$) support (${\tau }_{q_i}$), c) the determination of plesiomorphic (old) split-supporting site-patterns given two different polarities of character transformation along the internal branch of each possible quartet tree, ${\rho }_{\overrightarrow{q_{{}_1}}}$ and ${\rho }_{\overleftarrow{q_{{}_1}}}$, and d) estimation of expected convergent split-supporting site-patterns (${\kappa }_{\overrightarrow{q_{{}_1}}}$,${\kappa }_{\overleftarrow{q_{{}_1}}}$) supporting quartet q₁ in ML split pattern estimations using branch length and model optimization on constraint topologies of the other two possible quartet relationships (q₂, q₃).