Fig. 4

Evaluating and refining simian foamy virus (SFV)-host co-evolutionary model. a Calibrated maximum clade credibility Bayesian SFV phylogeny, estimated from an alignment of polymerase nucleotide sequences (74 sequences, 412 nt), using BEAST 1.8.0 [40]. The topology was fixed to the one we obtained from the MrBayes analyses. The calibrating nodes are encircled. The bars represent the uncertainty of the estimated node heights. The timescale is in millions of years. See Additional file 2: Table S1 for a complete list of species codes used in the study; PFV is primate foamy virus which is the new name given to HFV (human foamy virus). b Refined co-evolutionary history of SFVs (red) and their hosts (blue). Red dotted branch represents a possible Aotus ghost FV lineage that has not been sampled in our study. The directions of cross transmissions are indicated by arrows. Small arrows indicate cross-species transmission events, and large arrows indicate cross-genus transmission events. The red transparent bar shows the uncertainty of the cross-species transmission timing. Two alternative possible SFV cross-species transmission scenarios involving Cacajao, Cebus, and Callithrix monkeys are shown in green and blue. ‘Question marks’ indicate ambiguous cross-species transmission pathways. Co-speciation events at the genus level are indicated by solid red squares, and those at the species level are indicated by solid red circles. The trees are not scaled to time. c SFV-host divergence correlation analysis. Black dots represent co-diverging branches identified under the refined SFV-host co-evolutionary model (see panel b). A well supported linear correlation was found (linear regression: N = 16, R² = 0.8032, p < 0.0001), represented by a solid black line. The dots are labelled with roman numerals (I–XVI), referring to branches in Additional file 1: Figure S1