ribi: deadline Tuesday for NAEM conference

[richelbilderbeek]

To quote the NERN site:

Deadline for submissions of abstracts for a presentation in one of the parallel sessions: Tuesday 13 December 2016

[richelbilderbeek]

Perhaps a poster would be better, sent an email to @rsetienne.

[richelbilderbeek]

Contacted Boris Kramer, he said he'd place me iff I submit an abstract. Today's top priority job has shown itself!

[richelbilderbeek]

Sent in this:

The effect of excluding protracted speciation in phylogeny construction

We know that speciation takes time. This is suggested as one of the reasons why many species' lineages-through-time plots flatten off (do not increase as much) towards the present: the new species-to-be are not yet observed as being good species (Etienne & Rosindell, 2012).

Our current phylogeny inference tools ignore this fact: at a branching event both species are immediatly observed as good species, also when this takes place close to the present.

This research shows the effect of ignoring this fact.

This is done by first simulating phylogenies using the protracted birth-death model (Rosindell et al., 2010). From those fully-known phylogenies, a posterior is inferred using a standard birth-death model, by the Bayesian phylogeny inference tool BEAST2 (Bouckaert et al., 2014).

The original phylogeny and posterior are compared using the normalized lineages-through time statistic (Janzen et al., 2015), serving as a measure of error.

Using this error measure, it is shown when using the constant birth-death model suffices and when we should be more careful in drawing conclusions.

References

• Bouckaert, Remco, et al. "BEAST 2: a software platform for Bayesian evolutionary analysis." PLoS Comput Biol 10.4 (2014): e1003537.
• Etienne, Rampal S., and James Rosindell. "Prolonging the past counteracts the pull of the present: protracted speciation can explain observed slowdowns in diversification." Systematic Biology (2012): syr091.
• Janzen, Thijs, Sebastian Höhna, and Rampal S. Etienne. "Approximate Bayesian computation of diversification rates from molecular phylogenies: introducing a new efficient summary statistic, the nLTT." Methods in Ecology and Evolution 6.5 (2015): 566-575.
• Rosindell, James, et al. "Protracted speciation revitalizes the neutral theory of biodiversity." Ecology Letters 13.6 (2010): 716-727.

[richelbilderbeek]

Feedback from Boris Kramer:

i think you are over the official word count of 75 words (you can leave out references for an abstract). i guess it would be good to introduce with one sentence and point clearly to the weakness of the approaches taken so far. then followed by why and how you do it differently and what the results are. finishing sentence could be something like our approach shows that it is important to take this and that into account to overcome the weaknesses of the xx models. of course all this with approx 75 words....

one thing i im tempted to see in your abstract is that you didnt jet face the need to sell your work (yes, i dont like the idea either). if you need to publish you are forced to focus on whats new and whats interesting in which context for which reader. its really helpful to sometime gain a different and more focused and clear perspective on your own work.

[richelbilderbeek]

How would Boris like this...

Ignoring incipient species

The classic birth-death model incorrectly assumes that if a speciation event took place recently, the new species is directly recognized as such. This research is first in showing the consequence of this incorrect assumption. It uses simulating protracted phylogenies as a test case. From each of these, that phylogeny is inferred with a standard birth-death model, showing when the constant birth-death model suffices and when not.

[richelbilderbeek]

From Boris:

75 and it can be a bit more as the other ones also are ...

The classic birth-death model incorrectly assumes that if a speciation event took place recently, the new species is directly recognized as such and is then termed a good species. Here i (or we, its common to add your supervisor if you want, then use we) want to present the consequences of this often criticized and incorrect assumption (showing the consequence of this incorrect assumption). For this i sumulated.....It uses simulating protracted phylogenies as a test case. From each of these, that phylogeny is inferred with a standard birth-death model, showing when the constant birth-death model suffices and when not.... the results indicate that the constant birth death model....

( if you make the last two sentences a bit more clear then i like your abstract ;-) ) and a general statement that a non specielist on the topic can understand would also be nice ...

[richelbilderbeek]

I am now here:

Ignoring incipient species

The classic birth-death model incorrectly assumes that if a speciation event took place recently, the new species is directly recognized as a good species. Here I want to present the consequence of this often criticized and incorrect assumption. For this, I simulated protracted phylogenies and, via DNA sequences, infer phylogenies again using the standard birth-death model, showing when the constant birth-death model suffices and when not.

[richelbilderbeek]

Ignoring incipient species

The classic birth-death model incorrectly assumes that if a speciation event took place recently, the new species is directly recognized as a good species. Here I want to present the consequence of this often criticized and incorrect assumption. For this, I simulated protracted phylogenies and, via DNA sequences, infer phylogenies again using the standard birth-death model, showing when the constant birth-death model suffices and when not. The results indicate that the error made increases with the duration of speciation.

[richelbilderbeek]

@rsetienne will be happy that I've applied :sunglasses:

[richelbilderbeek]

From website: deadline poster: Tuesday 17 January 2017.

[richelbilderbeek]

From Boris:

the other organizer asked me if you could submit a longer abstract. i know its my fault and i told her that this is because i understood that the abstracts should be much shorter. so if you want to make yours a bit longer and detailed, the other organizer would be very happy.

[richelbilderbeek]

Ignoring incipient species

The classic birth-death model incorrectly assumes that if a speciation event took place recently, the new species is directly recognized as a good species.

Here I want to present the consequence of this often criticized and incorrect assumption.

To do so, phylogenies are simulated using a protracted birth-death model, to serve as 'the truth' in nature. DNA alignments true to these phylogenies are generated, to serve as simulated field data. From that data, it is tried to infer the original phylogeny back again, using (one of) the tools already present. The tool chosen is BEAST2, a Bayesian inference tool, for which a standard birth-death model tree prior was set up. Using BEAST2, from the simulated field data, a posterior is inferred and the error between original phylogeny and posterior is quantified with the normalized lineages-through-time statistic.

The results indicate that the error made increases with the duration of speciation.

[richelbilderbeek]

It has been done.

[richelbilderbeek]

NAEM 2017 Session 4A

December 20, 2016 5:08 pm 4 KB

your abstract has been selected for a presentation in session 4A (Eco-Evolutionary Theory)

cheers, Hanna ten Brink & Boris Kramer

Yes :smile: !