[Taxacom] Molecular dates are NOT necessarily minimums

[Michael Heads]

Hi Rob,

I'm not the only one to point out the 'problem of the priors'. Here's a
quick cut-and-paste of some material from 11 papers published after my 2012
review (sorry about the lack of proper formatting):

It is sometimes thought that ‘Fossils record the origins and disappearance
of organisms’ (Herrera, 2017, p. 2845), but in most cases a clade’s oldest
fossils post-date its actual time of origin by an unknown amount of time.
The priors used are crucial, but they are not observed or even calculated,
they are simply imposed; they represent “expert knowledge�, and these
“prior-encoded beliefs vary from expert to expert� (Landis, 2017, p. 129).
Experts in the Modern Synthesis tradition have assumed that a group is only
a little older than its oldest fossil, and modern Bayesian clock studies
maintain this view. However, there is no logical basis for it. How should
priors be selected? This is controversial, and “judgement of the degree to
which fossil minima approximate divergence timing . . . could be considered
a dark art . . . � (De Baets et al., 2016, p. 1). Naturally, the priors
that are specified have a great effect on the results. “Errors in the time
prior and in the rate prior can lead to very precise but grossly inaccurate
time estimates…’ (dos Reis et al., 2016, p. 74).

     Ho and Duchêne (2014) admitted that while the choice of priors ‘can
have a substantial impact on the estimates of node ages...’, it ‘is often
subjective and based on an overall interpretation of factors that are
difficult to quantify individually’. Hipsley and Müller (2014) also
admitted that selecting node age priors in fossil-based calibrations ‘is
often arbitrary or idiosyncratic at best...’; despite this, they supported
the method.

   Hauenschild et al. (2018) observed that vicariance is often ruled out
for clades in which reconstructed age estimates are younger than
continental breakup. However, they noted that the young ages estimated “may
also be due to overly small highest posterior density intervals of the time
estimated *based on too narrowly set priors**, *and vicariance thereby may
be erroneously ruled out (Beaulieu et al., 2013). If the importance of
vicariance in the course of the Gondwanan breakup is to be evaluated via
the pragmatic criterion of mismatch between age estimates and geological
history, the priors defining the molecular dating analyses have to be
solid� (italics added). Yet the priors used in practice are both subjective
and narrow, an inheritance from the earlier, simplistic idea that a group’s
age is the age of its oldest

fossil.

  ‘Errors in the time prior and in the rate prior can lead to *very precise
but grossly inaccurate *time estimates… ’ (dos Reis *et al*. 2016, p. 74).

   Ung, Michaux, & Leschen (2017) were sceptical about the practice of ‘ *.
. . *dating phylogenies using the age of the oldest fossil, which, despite
giving only a minimum age for divergence, becomes a maximum estimate by
proxy *. . . *’ (p. 433).

   Warnock et al. (2015) wrote that while: ‘Bayesian posterior estimates of
divergence times are extremely sensitive to the time priors’ (p. 1), ‘there
is frequently no material basis for selecting among the parameters’ (p. 6),
and the priors are ‘invariably established without justification’ (p. 2). Lee
and Skinner (2011: 540) agreed: ‘current practice often consists of little
more than educated guesswork’. Pirie and Doyle (2012: 108-109) concluded
that sporadic preservation in the fossil record ‘may be effectively
impossible to model’ and that priors ‘should . . . be regarded as highly
arbitrary’; their value ‘is something we simply cannot know’.

*References*

Beaulieu, J.M., Tank, D.C., & Donoghue, M.J. (2013). Southern Hemisphere
origin for campanulid angiosperms, with traces of the breakup of Gondwana. *BMC
Evol. Biol. *13: 80.

dos Reis M, Donoghue PC, Yang Z (2016) Bayesian molecular clock dating of
species divergences in the genomics era. *Nature Reviews. Genetics **17*,
71–80.

Landis, M.J., 2017. Biogeographic dating of speciation times using
paleogeographically informed processes. Syst. Biol. 66, 128–144.

Hauenschild, F., Favre, A., Michalak, I. & Muellner-Riehl, A.N. (2018). The
influence of the Gondwanan breakup on the biogeographic history of the
ziziphoids (Rhamnaceae). *J. Biogeogr*. 45: 2669–2677.

Herrera, J. P. (2017). Primate diversification inferred from phylogenies
and fossils. *Evolution **71*, 2845–2857.

Hipsley, C.A. and Müller, J., 2014. Beyond fossil calibrations: realities
of molecular clock practices in evolutionary biology. *Frontiers in
Genetics*, *5*, p.138.

Ho, S.Y. and Duchêne, S., 2014. Molecular�clock methods for estimating
evolutionary rates and timescales. *Molecular Ecology* *23: *5947-5965.

Landis, M.J., 2017. Biogeographic dating of speciation times using
paleogeographically informed processes. Syst. Biol. 66, 128–144.

Lee, Michael SY, and Adam Skinner. Testing fossil calibrations for
vertebrate molecular trees. *Zoologica Scripta* 40, no. 5 (2011): 538-543.

Pirie, M.D. and Doyle, J.A., 2012. Dating clades with fossils and
molecules: the case of Annonaceae. *Botanical Journal of the Linnean
Society*, *169*(1), pp.84-116.

Ung, V., Michaux, B. & Leschen, R. A. B. (2017). A comprehensive vicariant
model for Southwest Pacific biotas. *Australian Systematic Botany **29*,
424–439.

Warnock RCM, Parham JF, Joyce WG, Lyson TR, Donoghue PCJ (2015) Calibration
uncertainty in molecular dating analyses: there is no substitute for the
prior evaluation of time priors. *Proceedings of the Royal Society of
London – B. Biological Sciences **282*, 20141013.

On Mon, Oct 21, 2019 at 1:14 PM Rob Smissen via Taxacom <
taxacom at mailman.nhm.ku.edu> wrote:

> Once again Michael Heads promotes his self-serving slogan, claiming "As
> discussed, fossil-calibrated, molecular clock dates for clades can only
> ever be minimums."
> Current Bayesian molecular dating software gives a probability
> distribution of dates. How can a distribution of dates (a range) be a
> minimum? Well, Heads takes the oldest date from a published range and
> transmogrifies that into minimum.
>
> But that is only the beginning. The "only ever minimum" argument ignores
> the fact that other sources of error can contribute to misleadingly old
> date estimates. These include
>
>   1.  Ancestral polymorphism (splits in gene lineages are older than
> splits among populations). This has the opposite effect to the missing
> fossil problem. Gene lineage divergences necessarily predate population
> divergences, and therefore one might (wrongly) infer using Heads' logic
> that molecular date estimate are necessarily maximum.
>   2.  Rate variation among lineages.
>   3.  Miss-specified models.
>   4.  Incorrect attribution of fossils to extant lineage.
>   5.  Imprecise ages of fossils (often themselves given as ranges).
>   6.  Stochastic effects when sequences are short.
>
> Sometimes the data underpinning molecular dating papers is consistent with
> a 0 age (by which I mean that the samples could as well have been taken
> from a single panmictic population), yet Heads transmogrifies an age of
> millions of years into a so called minimum. I don't believe intelligent
> people would support that, at least not if they were honest.
>
> We should all take molecular clock dates with a big grain of salt, but
> slogans don't make for good science. I don't mind an argument that
> molecular date estimates may be underestimates, but I do mind the slogan
> that they are necessarily minimum. That is sophistry.
>
> Using the current generation of software, it is possible to explore the
> effect of using different priors. In fact it is quite easy. In particular,
> it is easy to explore the effect of increasing a calibration age to make
> dates older. It is also easy to constrain or relax the allowed rate
> variation among lineages.  If Heads was serious he could undertake these
> analyses. He could make the calibrations as old as he likes. Or stipulate
> that conspecific allopatric populations diverged in the Mesozoic as he oft
> claims. Then he could quantify how much rate variation, and how old the
> earth has to be, to accommodate his "biogeographic" assumptions. Despite
> claiming an understanding of molecular data and the workings of Bayesian
> analyses, and the fact that the software is free, Heads seems remarkably
> slow to provide analyses of how much older than their oldest fossils clades
> needs to be to fit his paradigm.
>
>
>
>
> ________________________________
>
> Please consider the environment before printing this email
> Warning: This electronic message together with any attachments is
> confidential. If you receive it in error: (i) you must not read, use,
> disclose, copy or retain it; (ii) please contact the sender immediately by
> reply email and then delete the emails.
> The views expressed in this email may not be those of Landcare Research
> New Zealand Limited. http://www.landcareresearch.co.nz
> _______________________________________________
> Taxacom Mailing List
>
> Send Taxacom mailing list submissions to: taxacom at mailman.nhm.ku.edu
> For list information; to subscribe or unsubscribe, visit:
> http://mailman.nhm.ku.edu/cgi-bin/mailman/listinfo/taxacom
> You can reach the person managing the list at:
> taxacom-owner at mailman.nhm.ku.edu
> The Taxacom email archive back to 1992 can be searched at:
> http://taxacom.markmail.org
>
> Nurturing nuance while assaulting ambiguity for 32 some years, 1987-2019.
>


-- 
Dunedin, New Zealand.

My books:

*Biogeography and evolution in New Zealand. *Taylor and Francis/CRC, Boca
Raton FL. 2017.
https://www.routledge.com/Biogeography-and-Evolution-in-New-Zealand/Heads/p/book/9781498751872


*Biogeography of Australasia:  A molecular analysis*. Cambridge University
Press, Cambridge. 2014. www.cambridge.org/9781107041028


*Molecular panbiogeography of the tropics. *University of California Press,
Berkeley. 2012. www.ucpress.edu/book.php?isbn—80520271968


*Panbiogeography: Tracking the history of life*. Oxford University Press,
New York. 1999. (With R. Craw and J. Grehan).
http://books.google.co.nz/books?id=Bm0_QQ3Z6GUC
<http://books.google.co.nz/books?id=Bm0_QQ3Z6GUC&dq=panbiogeography&source=gbs_navlinks_s>