The document describes the construction of networks to represent relationships between human genetic disorders and disease genes. Specifically, it details: 1) The creation of a "diseasome" bipartite network connecting 1,284 disorders and 1,777 disease genes based on known associations between genetic mutations and phenotypes. 2) The projection of this network into a "human disease network" where nodes represent disorders connected if they share disease genes, and a "disease gene network" where nodes represent genes connected if associated with the same disorder. 3) Analysis of the properties of these networks, finding most disorders are linked to only a few other disorders and disease genes, though some relate to dozens, and the networks display many connections

1. The Network of Driving Forces of
Global Environmental Change
Juan-Carlos Rocha, Oonsie Biggs & Garry Peterson
Stockholm Resilience Centre
Stockholm University

3. The challenge
Frequency and intensity of
regime shifts are likely to
increase.
ES’s may be substantially
affected.
Where?
Vulnerable areas?
Possible synergistic effects?
Cross-scale interactions?
Rockström et al., 2009

4. Regime shifts that matter to people
Regime shifts: Large, abrupt, persistent change in the structure and function of a
system.
Policy relevant = Substantial change in Ecosystem Services

5. Research agenda on RS: Early warnings!!
Bayesian
Web crawlers &
networks -
local knowledge
models
Knowledge of the
Models &
Jacobians
system
Statistics:
Autocorrelation
and variance
Data quality
(time series)

6. Research agenda on RS: Early warnings!!
Bayesian
Web crawlers &
networks -
local knowledge
models
Knowledge of the
Models &
Jacobians
system
? Statistics:
Autocorrelation
and variance
Data quality
(time series)

7. Virtruvian Man, Leonardo Da Vinci

8. Network Properties of Complex Human Disease Genes
The human disease network Identified through Genome-Wide Association Studies
´ ´ ´
Kwang-Il Goh*†‡§, Michael E. Cusick†‡¶, David Valleʈ, Barton Childsʈ, Marc Vidal†‡¶**, and Albert-Laszlo Barabasi*†‡**
Fredrik Barrenas1.*, Sreenivas Chavali1., Petter Holme2,3, Reza Mobini1, Mikael Benson1
*Center for Complex Network Research and Department of Physics, University of Notre Dame, Notre Dame, IN 46556; †Center for Cancer Systems Biology
(CCSB) and ¶Department of Cancer Biology, Dana–Farber Cancer Institute, 44 Binney Street, Boston, MA 02115; ‡Department of Genetics, Harvard Medical ˚ ˚
1 The Unit for Clinical Systems Biology, University of Gothenburg, Gothenburg, Sweden, 2 Department of Physics, Umea University, Umea, Sweden, 3 Department of
School, 77 Avenue Louis Pasteur, Boston, MA 02115; §Department of Physics, Korea University, Seoul 136-713, Korea; and ʈDepartment of Pediatrics and the Energy Science, Sungkyunkwan University, Suwon, Korea
McKusick–Nathans Institute of Genetic Medicine, Johns Hopkins University School of Medicine, Baltimore, MD 21205
Edited by H. Eugene Stanley, Boston University, Boston, MA, and approved April 3, 2007 (received for review February 14, 2007)
A network of disorders and disease genes linked by known disorder– known genetic disorders, whereas the other set corresponds to all Abstract
gene associations offers a platform to explore in a single graph- known disease genes in the human genome (Fig. 1). A disorder and
theoretic framework all known phenotype and disease gene associ- a gene are then connected by a link if mutations in that gene are Background: Previous studies of network properties of human disease genes have mainly focused on monogenic diseases
ations, indicating the common genetic origin of many diseases. Genes implicated in that disorder. The list of disorders, disease genes, and or cancers and have suffered from discovery bias. Here we investigated the network properties of complex disease genes
associated with similar disorders show both higher likelihood of associations between them was obtained from the Online Mende- identified by genome-wide association studies (GWAs), thereby eliminating discovery bias.
physical interactions between their products and higher expression lian Inheritance in Man (OMIM; ref. 18), a compendium of human
profiling similarity for their transcripts, supporting the existence of disease genes and phenotypes. As of December 2005, this list Principal findings: We derived a network of complex diseases (n = 54) and complex disease genes (n = 349) to explore the
distinct disease-specific functional modules. We find that essential contained 1,284 disorders and 1,777 disease genes. OMIM initially
human genes are likely to encode hub proteins and are expressed
shared genetic architecture of complex diseases. We evaluated the centrality measures of complex disease genes in
focused on monogenic disorders but in recent years has expanded comparison with essential and monogenic disease genes in the human interactome. The complex disease network showed
widely in most tissues. This suggests that disease genes also would to include complex traits and the associated genetic mutations that
play a central role in the human interactome. In contrast, we find that that diseases belonging to the same disease class do not always share common disease genes. A possible explanation could
confer susceptibility to these common disorders (18). Although this
the vast majority of disease genes are nonessential and show no history introduces some biases, and the disease gene record is far be that the variants with higher minor allele frequency and larger effect size identified using GWAs constitute disjoint parts
tendency to encode hub proteins, and their expression pattern indi- from complete, OMIM represents the most complete and up-to- of the allelic spectra of similar complex diseases. The complex disease gene network showed high modularity with the size
APPLIED PHYSICAL
cates that they are localized in the functional periphery of the of the largest component being smaller than expected from a randomized null-model. This is consistent with limited sharing
date repository of all known disease genes and the disorders they
SCIENCES
network. A selection-based model explains the observed difference
confer. We manually classified each disorder into one of 22 disorder of genes between diseases. Complex disease genes are less central than the essential and monogenic disease genes in the
between essential and disease genes and also suggests that diseases
classes based on the physiological system affected [see supporting human interactome. Genes associated with the same disease, compared to genes associated with different diseases, more
caused by somatic mutations should not be peripheral, a prediction
information (SI) Text, SI Fig. 5, and SI Table 1 for details]. often tend to share a protein-protein interaction and a Gene Ontology Biological Process.
we confirm for cancer genes.
Starting from the diseasome bipartite graph we generated two
biological networks ͉ complex networks ͉ human genetics ͉ systems
biologically relevant network projections (Fig. 1). In the ‘‘human Conclusions: This indicates that network neighbors of known disease genes form an important class of candidates for
biology ͉ diseasome
disease network’’ (HDN) nodes represent disorders, and two identifying novel genes for the same disease.
disorders are connected to each other if they share at least one gene
DISEASOME in which mutations are associated with both disorders (Figs. 1 and
D ecades-long efforts to map human disease loci, at first genet-
ically and later physically (1), followed by recent positional
cloning of many disease genes (2) and genome-wide association phenome
2a). In the ‘‘disease gene network’’ (DGN) nodes represent disease
genes, and two genes are connected if they are associated with the
Citation: Barrenas F, Chavali S, Holme P, Mobini R, Benson M (2009) Network Properties of Complex Human Disease Genes Identified through Genome-Wide
Association Studies. PLoS ONE 4(11): e8090. doi:10.1371/journal.pone.0008090
disease same disorder (Figs. 1 andgenome we discuss the potential of
disease 2b). Next,
studies (3), have generated an impressive list of disorder–gene Editor: Thomas Mailund, Aarhus University, Denmark
association Human Disease Network
these networks to help us understand andDiseasein a single represent Gene Network
pairs (4, 5). In addition, recent efforts to map Ataxia-telangiectasia the Received September 15, 2009; Accepted November 3, 2009; Published November 30, 2009
framework all known diseaseAR gene and phenotype associations.
protein–protein interactions in humans (6, 7), together with efforts hypospadias
(HDN)
to curate an extensive map of human metabolism (8) and regulatory insensitivity
Perineal
Androgen ATM
(DGN) Copyright: ß 2009 Barrenas et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
Properties of the HDN. If each human disorder tends to have a unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
networks offer increasingly detailed maps of the relationships T-cell lymphoblastic leukemia
Charcot-Marie-Tooth disease
between different disease genes. Most of the successful studies serous carcinoma Papillary
distinct and unique geneticBRCA1 then the HDN would be dis-
origin, HEXB Funding: This work was supported by the Swedish Research Council, The European Commission, The Swedish Foundation for Strategic Research (PH), and the
building on these new approaches have focused, however, on Prostate cancer
Lipodystrophy a connected into many single BRCA2 corresponding to specific disor-
nodes ALS2
LMNA
WCU (World Class University) program through the National Research Foundation of Korea funded by the Ministry of Education, Science and Technology R31-R31-
Spastic ataxia/paraplegia
single disease, using network-based tools tosyndromea better under-
Silver spastic paraplegia gain
ders or grouped into small clusters of a few closely related disorders.BSCL2
CDH1 2008-000-10029-0 (PH). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
standing of the relationship between the genes implicated in Ovarian cancer a In contrast, the obtained HDN displays many connections between VAPB
GARS Competing Interests: The authors have declared that no competing interests exist.
selected disorder (9).
Amyotrophic lateral sclerosis
Sandhoff disease both individual disorders and disorder classes (Fig. 2a). Of 1,284
GARS
Here we take a conceptually different approach, exploring disorders, 867 have at least HEXB link to other disorders, and 516
Lymphoma one * E-mail: fredrik.barrenas@gu.se
Spinal muscular atrophy
whether human genetic disorders and the corresponding disease disorders form a giant component, suggesting that the genetic
KRAS AR . These authors contributed equally to this work.
genes might be related to each Androgenat a higher level of cellular and cancer
other insensitivity origins of most diseases, to some extent, are shared with other
Breast
organismal organization. Support hypospadiasvalidity of this approach
Prostate cancer Perineal for the
diseases. The number of genes associated with a disorder, s, has a
LMNA ATM
BRCA2
is provided by examples of genetic disorders that arise from broad distribution (see SI Fig. 6a), indicating that most disorders
MSH2 BRIP1
Pancreatic cancer
mutations in more than a single gene (locus heterogeneity). For
Lymphoma
relate to a few disease genes,PIK3CA whereas a handful of phenotypes, such
BRCA1
Introduction human interactome. A more recent report that evaluated the
example, Zellweger syndrome is caused by mutations in any of at tumordeafness (s ϭ 41), leukemia (s ϭ 37), and colon cancer (s KRAS
Wilms tumor
Wilms as ϭ 34), network properties of disease genes showed that genes with
TP53
least 11 genes, all associated with peroxisome biogenesis Spinal muscular atrophyto dozens of genes (Fig. 2a). The degree (k) distribution of
Breast cancer
Ovarian cancer (10). relate RAD54L
TP53
Systems Biology based approaches of studying human genetic intermediate degrees (numbers of neighbors) were more likely to
Similarly, there are many examples of different mutations in the
Pancreatic cancer HDN (SI Fig. 6b) indicatesMAD1L1most disorders are linked to only
that diseases have brought in a shift in the paradigm of elucidating harbor germ-line disease mutations [12]. However, interpretation
same gene (allelic heterogeneity) giving rise to phenotypes cur- disease
Papillary serous carcinoma Sandhoff
RAD54L
MAD1L1 CHEK2 disease mechanisms from analyzing the effects of single genes to of this dataset might not be applicable to complex disease genes
Fanconi anemia
rently classified as different disorders. For example, mutations in
T-cell lymphoblastic leukemia
Lipodystrophy
PIK3CA understanding the effect of molecular interaction networks. Such
Author contributions: D.V., B.C., M.V., and A.-L.B. designed research; K.-I.G. and M.E.C.
VAPB since 97% of the disease genes were monogenic. Despite this
TP53 have been linked to 11 clinically distinguishable cancer- Charcot-Marie-Tooth disease
performed research; K.-I.G. and M.E.C. analyzed data; and K.-I.G., M.E.C., D.V., M.V., and
related disorders (11). Given the highly interlinked internalAmyotrophic lateral sclerosis the paper.
Ataxia-telangiectasia orga- A.-L.B. wrote CHEK2 CDH1 MSH2 networks have been exploited to find novel candidate genes, based reservation, both the latter studies found a functional clustering of
nization of the cell (12–17), it should be possible to improve the The authors declare no conflict of interest.
BSCL2
on the assumption that neighbors of a disease-causing gene in a disease genes. Another concern is that the above studies could be
single gene–single disorder approach by developing a conceptual Silver spastic paraplegia syndrome
This article is a PNAS Direct Submission. network are more likely to cause either the same or a similar confounded by discovery bias, in other words these disease genes
framework to link systematically all genetic disorders (the human ataxia/paraplegiaSpastic
ALS2
disease [1–14]. Initial studies investigating the network properties
Abbreviations: DGN, disease gene network; HDN, human disease network; GO, Gene were identified based on previous knowledge. By contrast,
‘‘disease phenome’’) with the complete list of disease genes (the anemia BRIP1
Fanconi Ontology; OMIM, Online Mendelian Inheritance in Man; PCC, Pearson correlation coeffi-
‘‘disease genome’’), resulting in a global view of the ‘‘diseasome,’’
of human disease genes were based on cancers and revealed that Genome Wide Association studies (GWAs) do not suffer from
cient.
the combined set of all known disorder/disease gene associations. **To whom correspondence may be addressed. E-mail: alb@nd.edu or marc࿝vidal@
up-regulated genes in cancerous tissues were central in the such bias [15].
dfci.harvard.edu. interactome and highly connected (often referred to as hubs) In this study, we have derived networks of complex diseases and
Results This article contains supporting information online at www.pnas.org/cgi/content/full/ [1,2]. A subsequent study based on the human disease network
Fig. 1. Construction of the diseasome bipartite network. (Center) A small0701361104/DC1.
subset of OMIM-based disorder– disease gene associations (18), where circles and rectangles complex disease genes to explore the shared genetic architecture of
Construction of the Diseasome. We constructed a bipartite graph
and disease gene network derived from the Online Mendelian complex diseases studied using GWAs. Further, we have evaluated
consisting ofto disorders and disease genes, respectively. A link isto all between aThe National Academy of Sciences ofif mutations in that gene lead to the specific disorder.
correspond
two disjoint sets of nodes. One set corresponds placed © 2007 by disorder and a disease gene the USA
The size of a circle is proportional to the number of genes participating in the corresponding disorder, and the color corresponds to the disorder class to which the disease Inheritance in Man (OMIM) demonstrated that the products of the topological and functional properties of complex disease genes
belongs. (Left) The HDN projection of the diseasome bipartite graph, in which two disorders are connected if there is a gene that is implicated in both. The width of disease genes tended (i) to have more interactions with each other in the human interactome by comparing them with essential,
www.pnas.org͞cgi͞doi͞10.1073͞pnas.0701361104 PNAS ͉ May 22, 2007 ͉ vol. 104 ͉ no. 21 ͉ 8685– 8690
a link is proportional to the number of genes that are implicated in both diseases. For example, three genes are implicated in both breast cancer and prostate cancer,
than with non-disease genes, (ii) to be expressed in the same tissues monogenic and non-disease genes. We observed that diseases
resulting in a link of weight three between them. (Right) The DGN projection where two genes are connected if they are involved in the same disorder. The width of
a link is proportional to the number of diseases with which the two genes are commonly associated. A full diseasome bipartite map is provided as SI Fig. 13.
and (iii) to share Gene Ontology (GO) terms [8]. Contradicting belonging to the same disease class do not always show a tendency
earlier reports, this latter study demonstrated that the non-essential to share common disease genes; the complex disease gene net-
human disease genes showed no tendency to encode hubs in the work shows high modularity comparable to that of the human
a few other disorders, whereas a few phenotypes such as colon mentary, gene-centered view of the diseasome. Given that the links
cancer (linked to k ϭ 50 other disorders) or breast cancer (k ϭ 30) signify related phenotypic association between two genes, they
represent hubs that are connected to a large number of distinct represent a measure of their phenotypic relatedness, which could be PLoS ONE | www.plosone.org 1 November 2009 | Volume 4 | Issue 11 | e8090
disorders. The prominence of cancer among the most connected used in future studies, in conjunction with protein–protein inter-
disorders arises in part from the many clinically distinct cancer actions (6, 7, 19), transcription factor-promoter interactions (20),
subtypes tightly connected with each other through common tumor and metabolic reactions (8), to discover novel genetic interactions.
repressor genes such as TP53 and PTEN. In the DGN, 1,377 of 1,777 disease genes are connected to other
Although the HDN layout was generated independently of any disease genes, and 903 genes belong to a giant component (Fig. 2b).
knowledge on disorder classes, the resulting network is naturally Whereas the number of genes involved in multiple diseases de-
and visibly clustered according to major disorder classes. Yet, there creases rapidly (SI Fig. 6d; light gray nodes in Fig. 2b), several
are visible differences between different classes of disorders. disease genes (e.g., TP53, PAX6) are involved in as many as 10
Whereas the large cancer cluster is tightly interconnected due to the disorders, representing major hubs in the network.
many genes associated with multiple types of cancer (TP53, KRAS,
ERBB2, NF1, etc.) and includes several diseases with strong pre- Functional Clustering of HDN and DGN. To probe how the topology

10. Regime shift database

11. Regime shift database

12. Regime shift database
Description of the alternative
regimes and reinforcing
feedbacks
The drivers that precipitate the
regime shift
Impacts on ecosystem services
and human well-being
Management options
www.regimeshifts.org

13. N Policy relevant regime shifts Mechanism Reversibility
1 Bivalves collapse Established H
2 Coral transitions Established H
3 Desertification Contested H, I
4 Encroachment Established H
5 Eutrophication Established H, I, R
6 Fisheries collapse Contested U
7 Marine foodwebs collapse Contested U
8 Forest - Savanna Established I
9 Hypoxia Established H, R
10 Kelp transitions Established H, R
11 Soil salinization Established H, I
12 Steppe - Tundra Established I
13 Tundra - Forest Established I
14 Monsoon circulation Established I
15 Thermohaline circulation collapse Established I
16 Greenland ice sheet collapse Established I
17 Arctic salt marshes Established I
18 Peatlands Established I
19 River channel position Established I
20 Soil structure Established H, I
Reversibility: H = Hysteretic; I = Irreversible; R= Reversible; U = Unknown
Current data: 20 Regime Shifts in Social-Ecological Systems

14. Hurricanes tides
Thermal anomalies in summerLow
Ocean acidification
Sea level rise
Disease
Fishing technology
Pollutants Wind stress
25
Thermal low pressure
Upwellings
Water column density contrast
Invasive species Sediments Tragedy of the commons
Urban storm water runoff
Fishing
Water vapor
Turbidity Urbanization Sea surface temperature
Sewage
20
Daily Relative cooling
Coral.transitions Logging
Salt.marshes Marine.foodwebs
Nutrients inputs
Fisheries.collapse house consumption preferences
Green Fish gases
Water stratification Kelps.transitions
Precipitation Bivalves.collapse
Number of vertex
15
River.channel.change Hypoxia
Floating.plants
Flushing Fertilizers use ENSO like events
Erosion Food supply
Eutrophication Subsidies
Floods Demand
Global warming
Impoundments Human population
Agriculture Access to markets
10
Deforestation
Leaking Termohaline.circulation
Forest.to.savannas
Rainfall variability
Landscape fragmentation Immigration
Greenland
Peatlands Monsoon.weakening
Soil.salinization
Irrigation
5
Encroachment Tundra.to.Forest
Dry.land.degradation Infrastructure development
Droughts
Migration
Aquifers
Drainage
Fire frequency Temperature Dry−spells
0
Atmospheric CO2
Irrigation infrastructure Soil.structure Managerial practices diversity
1 2 3 4 5 6 7 9 10 11 12 13 14 15 16 18 19 20 22 23 26
Ranching (livestock)
Water infrastructure
Degree
Water availability Development policies
Production intensification cycles
Length of production
Labor availability
Food prices
Regime Shifts - Drivers
Bipartite Network

15. Soil.structure
40
Dry.land.degradation
Soil.salinization
Peatlands
Fisheries.collapse
Salt.marshes
Bivalves.collapse
30
Encroachment
Number of links
Greenland Coral.transitions
Hypoxia
Eutrophication River.channel.ch
20
Forest.to.savannas
Kelps.transitions
Tundra.to.Forest
10
Floating.plants
Termohaline.circulation Monsoon.weakening
Marine.foodwebs
0
1 2 3 4 5 6 7 8 10 11 12 13 15 17
Number of Drivers shared
Regime Shifts Network

16. 500
400
Number of links
300
200
100
0
1 2 3 4 5 6 7 8 9 10 11
Number of Regime Shifts jointly caused
Drivers Network

17. 500
400
Number of links
300
200
100
0
1 2 3 4 5 6 7 8 9 10 11
Number of Regime Shifts jointly caused
Drivers Network

18. Green house gases
500
Global warming
400
Turbidity
Number of links
Fishing Food supply
300
Nutrients inputs Irrigation
200
Fertilizers use Agriculture
Human population
Demand
100
Sewage
Deforestation
Floods
0
1 2 3 4 5 6 7 8 9 10 11
Urbanization
Number of Regime Shifts jointly caused Erosion
Droughts
Drivers Network

19. How our results differ from random?
Average Degree in simulated DN Co−occurrence Index
2500
3000
2500
2000
2000
1500
Frequency
Frequency
1500
1000
1000
500
500
0
0
29 30 31 32 33 34 35 36 −1776.6 −1776.4 −1776.2 −1776.0 −1775.8
Mean Degree s−squared

20. Causal-loop diagrams is a
N Policy relevant Regime Shifts Mechanism Reversibility
technique to map out the
1 Bivalves collapse Established H feedback structure of a system
2 Coral transitions Established H (Sterman 2000)
3 Coral bleaching Established H
4 Desertification Contested H, I
5 Encroachment Established H
6 Eutrophication Established H, I, R
7 Fisheries collapse Contested U
8 Marine foodwebs collapse Contested U
9 Forest - Savanna Established I
10 Hypoxia Established H, R
11 Kelp transitions Established H, R
12 Soil salinization Established H, I
13 Steppe - Tundra Established I
14 Tundra - Forest Established I
15 Monsoon circulation Established I
16 Thermohaline circulation collapse Established I
17 Greenland ice sheet collapse Established I
18 Arctic salt marshes Established I
19 Arctic ice collapse Established I
Reversibility: H = Hysteretic; I = Irreversible; R= Reversible; U = Unknown
Current data: 19 Regime Shifts descriptions + CLD.

21. Topological features of Causal Network
Centrality Definition
Degree The number edges a vertex is connected to
(Newman 2010): In-degree and Out-degree
Betweenness The extent to which a vertex lies on paths
between other vertices (Newman 2010)
Eigenvector A vertex is important if it is directly or Degree centrality
indirectly connected to other vertices that are
in turn important (Allesina and Pascual 2009),
like Google PageRank

22. Topological features of Causal Network
Centrality Definition
Degree The number edges a vertex is connected to
(Newman 2010): In-degree and Out-degree
Betweenness The extent to which a vertex lies on paths
between other vertices (Newman 2010)
Eigenvector A vertex is important if it is directly or Betweenness centrality
indirectly connected to other vertices that are
in turn important (Allesina and Pascual 2009),
like Google PageRank

23. Topological features of Causal Network
Centrality Definition
Degree The number edges a vertex is connected to
(Newman 2010): In-degree and Out-degree
Betweenness The extent to which a vertex lies on paths
between other vertices (Newman 2010)
Eigenvector A vertex is important if it is directly or Eigenvector centrality
indirectly connected to other vertices that are
in turn important (Allesina and Pascual 2009),
like Google PageRank

24. D1
1. What are the major global change
drivers of regime shifts? RS1 RS2 RS3
80
60
Numbervertex vertex
Number vertexvertex
50
60
40
of
Number of of
Number of
40
30
20
20
10
0
0
1 2 3 4 5 6 7 8 9 11 12 14 15 17 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 19 22
Outgoing links
Outdegree
Incoming links
Indegree
Few nodes have a lot of links!

25. D1
Marine Regime Shifts RS1 RS2 RS3
Local centrality Global centrality
0.12
0.10
Nutrients input
10
Phytoplankton
Nutrients input
Fishing
0.08
Dissolved oxygenMid−predators
Noxious gases
Global warming
Betweenness
Algae Bivalves abundance
Outdegree
Agriculture Bivalves abundance
0.06
Floods Zooplankton
5
Top predators Space
GlobalUrban Macrophytes Phytoplankton
Planktivore fish
warminggrowth Dissolved oxygen
Turbidity
SST Erosion SST
ENSO−like Water temperature
events frequency
Canopy−forming algae algae
Turf−forming Biodiversity
Fishing
0.04
Greenhouse gasesand meso−predators
Disease outbreak Urchin barren
Lobsters Nekton Coral abundance
Unpalatability
AtmosphericDemand
Water vapor
CO2 Plankton and Macroalgae abundance
Human population Upwellings
ConsumptionFertilizers use runoff filamentous algae
Precipitation Flushing Coral abundance
Urban Sewage
Deforestation Sediments
preferences
Localstorm water Herbivores
Landscape fragmentation/conversion
water movements
Disease outbreak
Tragedy of thecolumn acidification
Impoundments densityLeakage
Water frequency
OceanIrrigation contrast
Thermal annomalies species
Invasive
Droughts
Perverse incentives mixing
TechnologyWater Zooxanthellae
Low tides commons
Sulfide stress
Wind release
Stratification relative cooling structural complexity
Mortality rate
Habitat
Density Thermal Fishmatter
Daily competitors
SubsidiesPollutants low pressurecolumn
Hurricanescontrast in the water
Noxious gases
Trade Other Organic Phosphorous in water Water vapor
0.02 Biodiversity Zooplankton
Nekton
Space Upwellings
0
Mid−predators
Turbidity Algae
Water temperature
Greenhouse gases Floods
Thermal low pressureErosion Macrophytes
Turf−forming algae
Macroalgae abundance Flushing
Lobsters and meso−predatorsTop predators
Wind stress
Water column density contrast
Urchin barren
Herbivores
Canopy−forming algae
Habitat structural complexity
Phosphorous in growth
Urban
Density contrast inOrganic matter and filamentous algae
Leakage Plankton
0.00
Zooxanthellae mixing water
ENSO−like events water column
Mortality the
Unpalatability frequency
Droughts
OceanHumanPerverseDemand
rate Agriculture Planktivore fish
AtmosphericWater Technology preferences
Landscape coolingwater incentives
fragmentation/conversion
acidification theuse
Other competitors Sediments
DailyInvasiveLocalSewage runoff
Low PollutantsFish Subsidies
population
HurricanesCO2 release
Consumption
relativePrecipitationTrade
Deforestation movements
Thermal annomalies of water
tidesUrban Stratificationcommons
storm
Fertilizers
Irrigation
frequency
Tragedy
Impoundments
species
Sulfide
0.00 0.02 0.04 0.06 0.08 0.10 0.12
0 5 10 15
Eigenvector
Indegree

26. D1
Terrestrial Regime Shifts RS1 RS2 RS3
Local centrality Global centrality
0.08
8
Fire frequency Precipitation
0.06
Global warming Precipitation Agriculture
Woody plants dominance
6
Fire frequency
Forest Grass dominance Deforestation
Cropland−Grassland area Deforestation
Betweenness
Outdegree
Agriculture Irrigation Albedo
0.04
Albedo Grass dominance
4
Irrigation
Rainfall variability
Soil productivity Forest
Droughts
DemandLand−Ocean temperature
Rainfall deficit
Savanna Native vegetation gradient
Woody plants dominance
Demand
Productivity
Land−Ocean temperature gradient
Atmospheric temperature
Erosion
Savanna
SST Atmospheric temperature
Floodsdemand
Grazing Water infrastructure Evapotranspiration
Water Erosion
Vegetation Space
Water availability
2
Atmospheric CO2
0.02
Human population Palatability
Soil moisture productivity
Soil Vegetation
Water infrastructure
Water availability
Advection
Carbon storage Global warming
Soil impermeability Solar radiation
Infrastructure developmentstress
WindTree release
maturity
Aquifers
LatentSoil quality
heatevents
Monsoon circulation
ENSO−likeDust frequency Vapor Soil salinity Soil salinity
Biomass
Logging industryShadow_rooting level
ImmigrationWater consumption
Land−Ocean pressure gradient concentration Productivity Aerosol concentration Soil moisture Rainfall deficit
use Moisture Carbon storage
Lifting Ranching
condensation Advection
FertilizersAbsorption of solar radiation
Aerosol Brown radiation
Solar clouds
Illegal logging
Sea tides Brown clouds Roughness
Temperature
Land conversion Ground water table
Grazers Absorption of solar radiation
Aquifers Evapotranspiration variability
Land conversion Rainfall Cropland−Grassland area
Vapor Droughts
Native vegetation
Ground Waterstress frequencyGrazers
ENSO−like events
SSTMonsoon
Land−Ocean water table
pressure gradient circulation
Wind demand
WaterTemperature
Shadow_rooting Moisture
Dust LiftingRoughnessTree maturity
Soil quality
consumptioncondensation level
Palatability
0
0.00
RanchingFloods
Grazing Space
Soil impermeabilityBiomass population
Human
Latent heat Logginglogging Atmospheric CO2
Fertilizers Illegal development
Immigration
Sea tides releaseindustry
Infrastructure
use
0 2 4 6 8 0.00 0.02 0.04 0.06 0.08
Indegree Eigenvector

27. Interaction of regime
shifts drivers?
Regime shifts are tightly connected. The
management of immediate causes or well
studied variables might not be enough to
avoid such catastrophes.
Agricultural processes and global warming
are the main causes of regime shifts.
Network analysis might be a useful
approach to address causality relationships

28. Thanks!
Drs. Oonsie Biggs & Garry
Peterson for their supervision
RSDB folks for inspiring
discussion and writing
examples
SRC for an inspiring research
space and funding!
Questions??
e-mail: juan.rocha@stockholmresilience.su.se
Twitter: @juanrocha
Blog: http://criticaltransitions.wordpress.com/
What is a regime shift?
Science pub May 2009 - SRC