Dating Tips for Divergence-Time Estimation. Issue 11 (November 2015)

• Record Type: Journal Article
• Authors: O'Reilly, Joseph E.; dos Reis, Mario; Donoghue, Philip C.J.
• Is Part Of: Trends in genetics. Volume 31:Issue 11(2015)
• Page Start: 637
• Page End: 650
• Publication Date: 2015-11
• Subjects: Genetics -- Periodicals; 576.5
• DOI: 10.1016/j.tig.2015.08.001
• Languages: English
• ISSNs: 0168-9525

Abstract

Abstract : The molecular clock is the only viable means of establishing an accurate timescale for Life on Earth, but it remains reliant on a capricious fossil record for calibration. 'Tip-dating' promises a conceptual advance, integrating fossil species among their living relatives using molecular/morphological datasets and evolutionary models. Fossil species of known age establish calibration directly, and their phylogenetic uncertainty is accommodated through the co-estimation of time and topology. However, challenges remain, including a dearth of effective models of morphological evolution, rate correlation, the non-random nature of missing characters in fossil data, and, most importantly, accommodating uncertainty in fossil age. We show uncertainty in fossil-dating propagates to divergence-time estimates, yielding estimates that are older and less precise than those based on traditional node calibration. Ultimately, node and tip calibrations are not mutually incompatible and may be integrated to achieve more accurate and precise evolutionary timescales. Trends: Total evidence dating constitutes a significant advance in divergence-time estimation. It overcomes problems with calibration by including fossil species on a par with their living relatives, using molecular sequence data from living species supplemented by morphological data from both living and fossil species. The method relies on the controversial hypothesis of a morphological clock and suffers from the lack of development of realistic models of morphological evolution. Most studies have failed to accommodate fossil age uncertainty. We present a protocol for characterizing and implementing this uncertainty, and demonstrate its impact on divergence-time estimation. We argue that total evidence dating encompasses a suite of methods that can be used in bespoke combinations chosen to best suit the nature of specific divergence-time estimation studies.