Presentation given to to Duke Information Future committee, March 2012

1. OPEN TREE OF LIFE
A community-assembled, continually updated evolutionary
history of all life
Karen A. Cranston
National Evolutionary Synthesis Center
Duke University

2. Tree of life
• 1.8million named
species
• Millions
more
unnamed / undiscovered
All biodiversity linked through common ancestry; tree of life is a grand challenge; scale is
huge

3. Correspondence and requests for materials should be addressed to C.W. (e-mail: whitlock@oregon. and elephant shrews with paenungulates. What is most unexpected
uoregon.edu). is that golden moles, a family of insectivores, are also part of this
clade. 12S rRNA sequences earlier suggested an association of
golden moles with paenungulates, but did not provide convincing
bootstrap support for this hypothesis7. Our expanded data set
Endemic African mammals demonstrates that insectivores are not monophyletic (Table 2)
shake the phylogenetic tree
Table 1 Bootstrap support for select clades based on different methods
Mark S. Springer*, Gregory C. Cleven*, Ole Madsen†,
Wilfried W. de Jong†‡, Victor G. Waddell§, Clade
Heather M. Amrine* & Michael J. Stanhope§ Paenungulata Paenungulata aardvark
elephant shrew
* Department of Biology, University of California, Riverside, California 92521, golden mole
USA Mitochondrial DNA
† Department of Biochemistry, University of Nijmegen, PO Box 9101, Parsimony 99 95
6500 HB Nijmegen, The Netherlands Transversion parsimony 64 90
Minimum evolution
‡ Institute for Systematics and Population Biology, University of Amsterdam, PO
Tamura–Nei I 100 92
Box 94766, 1090GT Amsterdam, The Netherlands Tamura–Nei II 100 78
§ Biology and Biochemistry, Queen’s University, 97 Lisburn Road, Logdet 99 90
Belfast BT9 07BL, UK Maximum likelihood 100 100
......................................................................................................................... vWF
Parsimony
The order Insectivora, including living taxa (lipotyphlans) and All positions 49 99
archaic fossil forms, is central to the question of higher-level 1st and 2nd positions 24 65
relationships among placental mammals1. Beginning with 3rd positions 51 93
Transversion parsimony 30 95
Huxley2, it has been argued that insectivores retain many primi- Minimum evolution
tive features and are closer to the ancestral stock of mammals than Tamura–Nei I 37 99
are other living groups3. Nevertheless, cladistic analysis suggests Tamura–Nei II 30 99
Logdet 43 97
that living insectivores, at least, are united by derived anatomical Maximum likelihood 78 100
features4. Here we analyse DNA sequences from three mito- A2AB
chondrial genes and two nuclear genes to examine relationships Parsimony
All sites 71 88
of insectivores to other mammals. The representative insectivores 1st and 2nd positions 49 81
are not monophyletic in any of our analyses. Rather, golden moles 3rd positions 31 67
are included in a clade that contains hyraxes, manatees, elephants,
Hedges, B. PNAS 98 (2001) Transversion parsimony 71 54
Minimum evolution
elephant shrews and aardvarks. Members of this group are of Tamura–Nei I 83 84
presumed African origin5,6. This implies that there was an exten- Tamura–Nei II 28 25
sive African radiation from a single common ancestor that gave Logdet 79 78
Maximum likelihood 81 89
rise to ecologically divergent adaptive types. 12S ribosomal RNA .............................................................................................................................................................................
Only two of the three paenungulate orders were represented among the mitochondrial and
transversions suggest that the base of this radiation occurred A2AB sequences. Tamura–Nei27 I and II distances were calculated by using an equal-rates
during Africa’s window of isolation in the Cretaceous period assumption and a gamma-distribution of rates, respectively.
Nature © Macmillan Publishers Ltd 1997
NATURE | VOL 388 | 3 JULY 1997 61
Springer, M.S. et al. Nature (1997)
Stanhope, M.J. et al. PNAS (1998)
Learned amazing things about biodiversity from studying the tree; radically changed views

4. EVOLUTION OF HUMAN PATHOGENS
Smith et al, Nature, 2009
Importance to human health and disease

5. MODERN BIOLOGY NEEDS TREES
Conventional Evolutionary
statistics assume: trees provide:
Modified from Garland and Carter, 1994

6. Phylogeny'papers,'1978;2008'
12000"
10000"
Number'of'papers'published'
8000"
Rapid"increase"in"applica?ons"of"
phylogeny,"beginning"in"early"1990s"
6000"
4000"
2000"
0"
1978" 979" 980"1981" 982" 983" 984"1985" 986"1987" 988" 989" 990"1991" 992" 993" 994"1995" 996"1997" 998" 999" 000"2001" 002" 003" 004"2005" 006"2007" 008"
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2
Year'
Source:"ISI"Web"of"Science""
We’ve been extremely productive

7. Where can I browse,
search and download a
complete tree of life?
You can’t. (Yet)

8. Phylogeny'papers,'1978;2008'
12000"
10000"
Number'of'papers'published'
8000"
Rapid"increase"in"applica?ons"of"
phylogeny,"beginning"in"early"1990s"
6000"
4000"
2000"
0"
1978" 979" 980"1981" 982" 983" 984"1985" 986"1987" 988" 989" 990"1991" 992" 993" 994"1995" 996"1997" 998" 999" 000"2001" 002" 003" 004"2005" 006"2007" 008"
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2
Year'
Source:"ISI"Web"of"Science""
It’s not because we haven’t been productive

9. Reason 1: silos of phylogenetic knowledge

10. DATA AVAILABILITY
High archival rate of sequence data
~4% of all published
phylogenetic trees
Reason 2: can’t do synthesis without data; community norm to deposit sequences but not
trees

11. Most trees published
as (beautiful) figures
in PDF files
EVOLUTION
not reusable!
Weigmann et al. PNAS, 2011
Fig. 1. Combined molecular phylogenetic tree for Diptera. Partitioned ML analysis of combined taxon sets of tier 1 and tier 2 FLYTREE data samples (−lnL =
344155.6169) calculated in RAxML. Circles indicate bootstrap support >80% (black/bp = 95–100%, gray/bp = 88–94%, white/bp = 80–88%). Nodes with im-
proved bootstrap values resulting from postanalysis pruning of unstable taxa are marked by stars (black/bp = 95–100%, gray/bp = 88–94%, white/bp = 80–
88%). Colored squares on terminal branches indicate the presence, in at least one species of a family, of ecological traits as shown to lower left. The number
Most phylogenetic knowledge locked in PDFs; even best case (here): trees and alignments in
of origins of each trait was estimated with reference to the phylogeny, the distribution of each trait among genera within a family, and the known biology of
the organisms.
treebase, sequences in GenBank but so much more data in this figure!

12. 1. Build and make publicly available the first complete draft
tree of life
2. Engage the community in refinement and annotation of
the draft tree
3. Promote a culture of data sharing among systematists
through software products that target key bottlenecks in
phylogenetic workflows
4. Develop, disseminate, and evaluate novel methods for
phylogenetic synthesis

13. + NSF Assembling the Tree of Life projects
+ recent high-profile phylogenies
+ ribosomal RNA trees for Bacteria and Archaea
+ unnamed environmental sequences
+ taxonomies of living and extinct species
Draft Tree of Life as seed for
collaboration
Get all data available! Part of draft tree will be good; parts not so good

14. Tools to browse,
update, monitor the
draft tree of life
Tools that incentivize
collaboration
Tree image modified from Tree of Life Web Project page http://tolweb.org/Nymphalidae/12172 Pictures by Katja Schulz (queen butterfly; CCAttribution-NonCommercial) and Charles Lam (via Flicker;CCAttribution-ShareAlike)
publish refinement tools along with draft tree; invite comment / critique / updating

15. SMART GENERATION OF FIGURES FOR
PUBLICATION
• Semantic annotation layers
• Collaborative editing
EVOLUTION
• Integrated submission of
data and annotations to
archives
ig. 1. Combined molecular phylogenetic tree for Diptera. Partitioned ML analysis of combined taxon sets of tier 1 and tier 2 FLYTREE data samples (−lnL =
44155.6169) calculated in RAxML. Circles indicate bootstrap support >80% (black/bp = 95–100%, gray/bp = 88–94%, white/bp = 80–88%). Nodes with im-
Motivated to organize data at this point in pipeline; want to capture digitally at this point
roved bootstrap values resulting from postanalysis pruning of unstable taxa are marked by stars (black/bp = 95–100%, gray/bp = 88–94%, white/bp = 80–
8%). Colored squares on terminal branches indicate the presence, in at least one species of a family, of ecological traits as shown to lower left. The number
f origins of each trait was estimated with reference to the phylogeny, the distribution of each trait among genera within a family, and the known biology of
he organisms.

16. AUTOMATIC UPDATING
update trees
with new
sequence data
detect and synthesize newly
published trees
Can’t rely on manual refinement alone; need automate based on new data (and need new
analytical method development)

17. community
refinement
ng the major evolutionary relationships recovered in this analysis of Carnivora. Illustrations of
om top): Nandinia binotata; Felidae (Lynx rufus); Viverridae (Viverra zibetha); Hyaenidae (Crocuta
y carnivorans (Eupleres goudotii); Canidae (Canis lupus); Ursidae (Ursus americanus); Phocidae
Odobenidae (Odobenus rosmarus); Ailurus fulgens; Mephitidae (Mephitis mephitis); Procyonidae
education and
outreach
(generalized schematic representing diverse taxa [African polecat and striped marten, badger,
telidae, Mustela (Mustela frenata); Mustelidae, Lutrinae (Lontra canadensis).
the major evolutionary relationships recoveredin this analysis of Carnivora. Illustrations of
nting the major evolutionary relationships recovered in this analysis of Carnivora. Illustrations of
top): Nandinia binotata; Felidae (Lynx rufus); Viverridae (Viverra zibetha); Hyaenidae (Crocuta
(from top): Nandinia binotata; Felidae (Lynx rufus); Viverridae (Viverra zibetha); Hyaenidae (Crocuta
gasy carnivorans (Eupleres goudotii); Canidae (Canis lupus); Ursidae (Ursus americanus); Phocidae
rnivorans (Eupleres goudotii); Canidae (Canis lupus); Ursidae (Ursus americanus); Phocidae
us); Odobenidae (Odobenus rosmarus); Ailurus fulgens; Mephitidae (Mephitis mephitis); Procyonidae
dobenidae (Odobenus rosmarus); Ailurus fulgens; Mephitidae (Mephitis mephitis); badger,
ids (generalized schematic representing diverse taxa [African polecat and striped marten, Procyonidae
Dynamic,
Mustelidae, schematic representing diverse Lutrinae (Lontra canadensis).
neralized Mustela (Mustela frenata); Mustelidae,taxa [African polecat and striped marten, badger,
dae, Mustela (Mustela frenata); Mustelidae, Lutrinae (Lontra canadensis).
continually updated
cladogram representing the major evolutionary relationships recovered in this analysis of Carnivora. Illustrations of
or lineages include (from top): Nandinia binotata; Felidae (Lynx rufus); Viverridae (Viverra zibetha); Hyaenidae (Crocuta
gos mungo); Malagasy carnivorans (Eupleres goudotii); Canidae (Canis lupus); Ursidae (Ursus americanus); Phocidae
tree of all life
Zalophus californianus); Odobenidae (Odobenus rosmarus); Ailurus fulgens; Mephitidae (Mephitis mephitis); Procyonidae
asal/other mustelids (generalized schematic representing diverse taxa [African polecat and striped marten, badger,
?
roup (Gulo gulo); Mustelidae, Mustela (Mustela frenata); Mustelidae, Lutrinae (Lontra canadensis).
?
?
automatic ?
updating
downstream
analyses
The grand vision

18. opentreeoflife.com
Karen Cranston, lead PI (Duke) Laura Katz (Smith)
Gordon Burleigh (Florida) Rick Ree (FMNH)
Keith Crandall (BYU) Stephen Smith (Michigan)
Karl Gude (MSU) Doug Soltis (Florida)
David Hibbett (Clark) Tiffani Williams (TAMU)
Mark Holder (Kansas)
AVAToL: Assembling, Visualizing and Analysis of
the Tree of Life, $5.76 million
The team and funding; empirical + computational biologists + comp sci + graphics /
communication