Origin and diversification of living cycads: a cautionary tale on the impact of the branching process prior in Bayesian molecular dating

Table 1 Age estimates for the six nodes subtending a fossil calibration (FC)

Fossil calibration strategy	Prior	HME	MLE (SS)	ESS (logL)	BF	Cycad stem (FC1)	*Dioon* stem (FC2)	*Bowenia* stem (FC3)	*Lepidozamia* stem (FC4)	Cycad crown (FC5)	Encephalarteae stem (FC6)
Fossil calibration strategy	Prior	HME	MLE (SS)	ESS (logL)	BF	min. age 265.1 Ma	min. age 56 Ma	min. age 33.9 Ma	min. age 33.9 Ma	min. age 235 Ma	min. age 72.1 Ma
Traditional fossil dataset	Birth-death	−26343.61	−26828.83	2451	26.7	316.6 [269.5-363.3]	140 [90-192.9]	116.3 [76.7-160.8]	38.1 [33.9-50.9]	257.2 [184.3-337.5]	129.2 [84.3-180.2]
Traditional fossil dataset	Yule	−26330.62	−26855.49	3835	26.7	295.7 [265.1-350.1]	206.4 [161.2-258.2]	170.3 [130.8-216.3]	68 [41.7-100.4]	277.2 [231-344.9]	190.4 [146.9-240.2]
New fossil dataset	Birth-death	−26343.36	−26825.04	2076	36.3	323.2 [273.9-364.6]	156.1 [107-207.9]	129.7 [88.7-174.3]	39.1 [33.9-55]	274.5 [235-332.4]	144.5 [97.7-192.5]
New fossil dataset	Yule	−26331.16	−26861.34	1361	36.3	296.9 [265.1-351.7]	203.5 [158.6-253.6]	169.5 [129.4-214.2]	67.8 [40.6-102]	280.1 [235.6-334.8]	188 [143.6-235.6]

Absolute ages and their credibility intervals were obtained for the two pairs of Bayesian calibration strategies, using either the Yule or the birth-death model as branching process prior.
Values are the median ages and 95% Highest Posterior Density (HPD). We denote HME: harmonic mean estimate; MLE (SS): marginal likelihood, estimated with the stepping-stone sampling; ESS: effective sample size; BF: Bayes factor (BF), computed as the difference between the MLE of the birth-death and the Yule prior. A positive value indicates support for the birth-death prior.

ISSN: 2730-7182