FAQ
What can I do with DateLife?
You can interact with it as a website to get estimates for the ages of various groups. You can also run automated queries against it. While still in alpha form, the api may change, but you can see what it is at any point by just looking at the URL. For example, searching for a Newick tree for three particular taxa, using only peer-reviewed studies, sends you to this page:
http://datelife.org/cgi-bin/R/result?input=Rhinoceros_unicornis%2CEquus_caballus%2CMus_musculus&format=html&partial=liberal&useembargoed=no&uncertainty=100&usetnrs=no&tnrssource=NCBI
You could have a python, perl, or R script call this same URL to get trees back. By changing format=newickmed to format=bestguess you get a single point estimate of the age of the MRCA of the group of taxa, which can be used by other scripts. It's fine to use automated tools to query the site, we just ask that you try not to hit it so frequently as to degrade performance for others (perhaps give it a few seconds between calls, schedule hard use for the middle of the night, etc.). We haven't yet tested at what usage frequency it stops working well. You can download the source code and run it locally, perhaps hacking it for some other purpose. If you reuse in that way, we'd appreciate being cited (citation being the currency of science). If you have an improvement to make, either send us a patch or sign up as a developer (though forking the code is also possible, of course). The main code is R, which is fairly easy to use. We're planning to turn much of the code into an open source standalone R package as well (while also keeping the rest open).
Can I watch a movie about it?
Here is a presentation we gave at iEvoBio in July, 2012:
What is next?
We are building up the back end database (please send us your published chronograms). We are also investigating incorporating fossil information.
What do I cite?
First, the papers for the trees and fossils you are using, if possible. It is like citing papers contributing data to GenBank: if you were to analyze one or two datasets, you would cite the papers that generated them; if you were to analyze all the plant rbcl sequences in GenBank, you might not. We also ask that you cite our website (we hope to have a peer-reviewed publication soon to cite instead). If you use the congruification method, you should also cite:
"Congruification: support for time scaling large phylogenetic trees" Eastman, Jonathan M; Harmon, Luke J; Tank, David C (2013) Methods in Ecology and Evolution, In press
Who is "we"?
Currently, Brian O'Meara, Jonathan Eastman, Tracy Heath, April Wright, Klaus Schliep, Scott Chamberlain, Peter Midford, Luke Harmon, Joseph Brown, Matt Pennell, Mike Alfaro. Maybe you, too?
Who paid for this?
DateLife was developed as part of the Phylotastic hackathon (parts I and II, at NESCent and iPlant, respectively), sponsored by NESCent, iPlant, and BioSynC. NESCent is funded by the National Science Foundation (Grant #EF-0905606) and your donations. iPlant is funded by a grant from the National Science Foundation Plant Cyberinfrastructure Program (#DBI-0735191). BioSynC is funded by the MacArthur Foundation as well as the National Science Foundation, The J.B. and M.K. Pritzker Family Foundation, The Harris Family foundation and MaryEllen and Richard Keyser, and many more donors. The computing infrastructure behind DateLife was funded by startup funds from the U. of Tennessee, Knoxville for Brian O'Meara.
What are the problems with DateLife?
The first comes from its very concept: using chronograms to date other chronograms (or sets of taxa). Eventually this can get you far from the original data used, often fossil calibrations, and so errors can be magnified. Trees in the database may not be independent for the same reasons. We suspect that, even with these issues, using chronograms from DateLife may still be better than making up branch lengths (Grafen or uniform) for later analyses, and this is also useful for trees for teaching, but it will nearly always be better to use actual calibrations and chronogram inference. Waiting to publish an analysis until you create your own well-calibrated tree is justifiably the gold standard, and for many researchers, also the minimal standard. There are myriad other issues, such as poor or missing coverage for some groups. There may also be undiscovered errors, or at least biases, in the algorithms used. If you do find any, please let us know; even better, write a patch to fix the problem.
Don't you know about TimeTree? Why do a new site?
TimeTree (Hedges et al., 2006 [PDF]) is a good resource for date information. It has expert-derived dates as well as dates from published trees. It also has a good web interface (for mobile devices, too) and now free iPhone and iPad apps. You can even get free book chapters and poster images or buy a physical book or poster. These are all remarkable tools for outreach and for answering targeted questions. What you may not do is reuse the data in an automated manner (at least when we started our project in June 2012). For example, TimeTree displays this note (as of June 8, 2012, checked again Jan. 30, 2013):
and the FAQ states (as of June 8, 2012, checked again Jan. 30, 2013):
Requests to allow download of the calibrated tree or other widespread reuse have been consistently but politely declined. Reusing existing work makes a great deal of sense, but since this was not possible, we decided to develop our own. It differs from TimeTree in other ways that we hope are improvements, such as allowing for information about uncertainty from individual studies, but the prohibition on large-scale reuse was really the key issue driving our development.
Who is NodeThirtyThree Design?
We have no idea. They make free web design templates with generous reuse licensing but ask for people to cite them. So we have.
Note source code is available at https://github.com/phylotastic/datelife (the R package with this functionality), https://github.com/phylotastic/datelifeweb (the package that generates the web page), and https://github.com/phylotastic/datelifedocker (the docker file to create the image).
DateLife is part of the Phylotastic project, funded by NSF. It uses resources from the Open Tree of Life project, rOpenSci, and TreeBase. To help, please add trees to OpenTree's database! [they can be yours or others']
