Advertisement
Research Article

The Effect of Inappropriate Calibration: Three Case Studies in Molecular Ecology

  • Simon Y. W. Ho mail,

    *E-mail: drsimonho@gmail.com

    Affiliation: Department of Zoology, University of Oxford, Oxford, United Kingdom

    Current address: Centre for Macroecology and Macroevolution, School of Botany and Zoology, Australian National University, Canberra, Australian Capital Territory, Australia

    X
  • Urmas Saarma,

    Affiliations: Department of Zoology, Institute of Ecology and Earth Sciences, University of Tartu, Tartu, Estonia, Estonian Biocentre, Tartu, Estonia

    X
  • Ross Barnett,

    Affiliation: Department of Biology, University of York, York, United Kingdom

    X
  • James Haile,

    Affiliation: Department of Zoology, University of Oxford, Oxford, United Kingdom

    X
  • Beth Shapiro

    Affiliation: Department of Zoology, University of Oxford, Oxford, United Kingdom

    X
  • Published: February 20, 2008
  • DOI: 10.1371/journal.pone.0001615
  • Published in PLOS ONE

Reader Comments (2)

Post a new comment on this article

Referee comments: Referee 2

Posted by PLoS_ONE_Group on 25 Feb 2008 at 10:58 GMT

Referee 2's review:

This manuscript presents three case studies as a mean to exemplify the impact of using inappropriate calibrations in molecular ecology studies. The authors built upon some of their earlier work showing that there is a time-dependency of evolutionary rate estimates. The authors argue that using an external interspecific calibration point when inferring intraspecific divergence dates scaling on genealogical times will significantly alter the results. They used three cases studies for illustrating the effects of (1) applying an inappropriate external rate from the literature, and (2) using an external calibration point. They conclude that the evolutionary inferences drawn in each case study are significantly altered by using revised internally-calibrated substitution rates. They also advocate the inclusion of ancient DNA sequences to obtain reliable estimates of intraspecific rates.

It is certainly of importance to point out the difficulties in estimating accurate intraspecific divergence dates given the current widespread use of both an ad-hoc substitution rate from the literature and inappropriate external calibrations. Therefore, this paper will likely be of interest to people involved in molecular ecology, phylogeography and also ancient DNA analyses.

However, I found the first example on birds quite trivial as it is just the demonstration that using a higher rate will result in younger estimates. I agree that there are several reasons why the use of the 'traditional' mitochondrial rate of 0.01 subs/site/Myr is questionable. However, I cannot see in the current manuscript any justification of using a revised rate of 0.075 subs/site/Myr (where does it come from?). Moreover, the particular problem of using a universal substitution rate in birds has already been pointed out and discussed in details by one of the authors in a specialized journal (Ho SYW. Journal of Avian Biology, 38: 409-414). Therefore, I found the case study of Pleistocene speciation in birds presented here of only limited interest.

I am more convinced by the other two other examples showing how the inclusion of ancient DNA sequences might help in estimating more accurate substitution rate and also demographic parameters as compared to the current practise of using an external and distant calibration point. However, what is really regrettable to judge of the real impact of such a practise is that clearly the analyses performed here cannot be replicated based on the limited technical information provided both in the main text and in the Materials and Methods section. Amazingly, there is no Materials and Methods section corresponding to the Bowhead Whale example which is perhaps the one for which is the most needed. We don't even know which ancient DNA sequences have been used for estimating the revised substitution rate. More generally much more details on both the dataset assembly and the analyses performed should be given in the Materials and Methods section which is currently inappropriate. Much more technical details should be provided for the analyses to be replicated.

Perhaps my main problem with the present manuscript is the absence of discussion of the critical opinion on the reality of the time-dependency of rate estimates expressed by Brent Emerson in its recent Systematic Biology paper (Syst. Biol. 2007, 56:337-345). This critical paper is not even cited in the present manuscript. The authors should at least take these counter arguments into account and discuss how they relate to the conclusions drawn in the present manuscript.

Finally, I found the Concluding remarks section of the manuscript particularly frustrating as it is currently not much more than a reassertion of what has been repeatedly pointed out in the main text. I think that it might rather be a good place for the authors to really take position and provide some guidelines for others to carry similar analyses following a (detailed) methodology they deem appropriate. Is it to say that researchers interested in estimating intraspecific divergence times should turn to ancient DNA in order to incorporate radiocarbon dated individuals?

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication the manuscript has been revised in light of these comments and to address other editorial requirements.


RE: Referee comments: Referee 2

simonywho replied to PLoS_ONE_Group on 26 Feb 2008 at 03:45 GMT

This is our response to the comments by Referee 2, submitted to PLoS ONE along with our revised manuscript:

> However, I found the first example on birds quite trivial as it is just the
> demonstration that using a higher rate will result in younger estimates. I agree
> that there are several reasons why the use of the 'traditional' mitochondrial rate
> of 0.01 subs/site/Myr is questionable. However, I cannot see in the current
> manuscript any justification of using a revised rate of 0.075 subs/site/Myr
> (where does it come from?). Moreover, the particular problem of using a
> universal substitution rate in birds has already been pointed out and discussed
> in details by one of the authors in a specialized journal (Ho SYW. Journal of
> Avian Biology, 38: 409-414). Therefore, I found the case study of Pleistocene
> speciation in birds presented here of only limited interest.

Response: We understand that our methodology of simply substituting a higher rate for a lower one may seem trivial at first glance, but we believe that the implications are far from trivial. As we have pointed out in the paper, all previous tests of the ‘Late Pleistocene Origins’ hypothesis have assumed the ‘traditional’ rate of 0.01 subs/site/Myr. This includes a very recent (2007) publication in ‘Science’ by Weir and Schluter.
Brief justifications for using the revised rate of 0.075 subs/site/Myr have been added to the main text as well as the Materials and Methods. We have also added a caveat regarding the universal applicability of a single rate, as noted by the reviewer.

**********

> I am more convinced by the other two other examples showing how the
> inclusion of ancient DNA sequences might help in estimating more accurate
> substitution rate and also demographic parameters as compared to the current
> practise of using an external and distant calibration point. However, what is
> really regrettable to judge of the real impact of such a practise is that clearly the
> analyses performed here cannot be replicated based on the limited technical
> information provided both in the main text and in the Materials and Methods
> section. Amazingly, there is no Materials and Methods section corresponding to
> the Bowhead Whale example which is perhaps the one for which is the most
> needed. We don't even know which ancient DNA sequences have been used
> for estimating the revised substitution rate. More generally much more details
> on both the dataset assembly and the analyses performed should be given in
> the Materials and Methods section which is currently inappropriate. Much more
> technical details should be provided for the analyses to be replicated.

Response: We suspect that the reviewer has accidentally overlooked the Materials and Methods section in our original submission, which contained a relatively detailed description of the bowhead whale data set and the Bayesian phylogenetic analysis. To make this clearer, we have now included explicit references to the Methods section in the main body of the text.

**********

> Perhaps my main problem with the present manuscript is the absence of
> discussion of the critical opinion on the reality of the time-dependency of rate
> estimates expressed by Brent Emerson in its recent Systematic Biology paper
> (Syst. Biol. 2007, 56:337-345). This critical paper is not even cited in the present
> manuscript. The authors should at least take these counter arguments into
> account and discuss how they relate to the conclusions drawn in the present
> manuscript.

Response: We have now cited the criticism by Emerson (2007), as well as a separate criticism by Bandelt et al. (2006). We responded directly to Emerson’s critique in a reply published elsewhere (Ho et al., 2007, Syst. Biol.), while an indirect response to Bandelt et al.’s criticism is in press for the American Journal of Human Genetics (Endicott & Ho). We have added a brief note to the revised paper to explain that the issue of internal vs external calibration still stands even if one is skeptical of the hypothesis of ‘time dependency of rate estimates’. This point is highlighted by the results summarized in the new Table 3.

**********

> Finally, I found the Concluding remarks section of the manuscript particularly
> frustrating as it is currently not much more than a reassertion of what has been
> repeatedly pointed out in the main text. I think that it might rather be a good
> place for the authors to really take position and provide some guidelines for
> others to carry similar analyses following a (detailed) methodology they deem
> appropriate. Is it to say that researchers interested in estimating intraspecific
> divergence times should turn to ancient DNA in order to incorporate radiocarbon
> dated individuals?

We have now added a discussion of the options available for internal calibration, as well as an outline of the drawbacks of intraspecific calibration.

**********