Systems genetics: From GWAS to disease pathways☆
Marijke R. van der Sijdea,1, Aylwin Ngb,c,2,3, Jingyuan Fua,⁎,3
aUniversity of Groningen, University Medical Centre Groningen, Department of Genetics, The Netherlands
bCentre for Computational and Integrative Biology and Gastrointestinal Unit, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA
cBroad Institute of Harvard and Massachusetts Institute of Technology, Cambridge, MA, USA
a b s t r a c ta r t i c l e i n f o
Received 2 September 2013
Received in revised form 21 March 2014
Accepted 27 April 2014
Available online 2 May 2014
In the past decade, genome-wide association studies (GWAS) have successfully identified thousands of genetic
variants underlying susceptibility to complex diseases. However, the results from these studies oftendo not pro-
vide evidence on how the variants affect downstream pathways and lead to the disease. Therefore, in the post-
GWAS era thegreatest challengeliesincombining GWAS findingswith additional molecular data to functionally
characterizetheassociations.The advances invarious~omicstechniqueshavemade itpossibletoinvestigatethe
effect ofrisk variants onintermediatemolecularlevels, such asgene expression, methylation, protein abundance
or metabolite levels. Asdisease aetiologyiscomplex,nosinglemolecular analysisisexpectedtofully unravelthe
bine data from different molecular levels and can help construct the causal inference from genotype to
tems biology by integrating genotype information with various ~omics datasets as well as with environmental
and physiological variables. In this review, we describe this approach and discuss how it can help us unravel
the molecular mechanisms through which genetic variation causes disease. This article is part of a Special Issue
entitled: From Genome to Function.
© 2014 Elsevier B.V. All rights reserved.
Over the last 5–10 years enormous progress has been made in iden-
ent complex diseases and complex traits. Genome-wide association
studies (GWAS), which aim to correlate allele frequencies of single nu-
ics research . At the moment (March 2014), 1818 GWAS papers
have been published describing 12,498 associations, as listed in the
GWAS catalogue of the National Human Genome Research Institute
(www.genome.gov/gwastudies/) . And new loci are still being dis-
covered. From a medical perspective, the ultimate goal of GWAS is to
identify the causal variants of a phenotype, their functional effects and
pathogenesis. However, although thousands of phenotype-associated
loci have been identified, the molecular mechanisms through which
they act are still largely unknown. In practice, the interpretation of
GWAS findings is complicated by the fact that most identified associa-
tions are part of a larger region of correlated variants. SNPs in close
ing it hard to pinpoint the causal variant. Furthermore, the majority of
identified SNPs are annotated outside of protein-coding genes, indicat-
ing that the underlying mechanism is most likely regulatory. This
makes it more challenging to elucidate the direct functional conse-
quences of the variants.
The lack of explanatory power of GWAS thus calls for additional
methods to uncover the mechanisms that underlie complex diseases.
Over thelast few years, different approaches have been explored to dis-
cover functional relationships between genes at the associated loci, for
instance, by searching for genes with similar functions or within the
same molecular pathway [3–5]. One widely used approach is the gene
set enrichment analysis, which determines whether an a priori defined
set of genes is statistically enriched for disease associations. Many tools
Biochimica et Biophysica Acta 1842 (2014) 1903–1909
☆ This article is part of a Special Issue entitled: From Genome to Function.
⁎ Corresponding author at: University of Groningen, University Medical Centre
Groningen, Department of Genetics, PO Box 30.001, 9700 RB Groningen, The
Netherlands. Tel.: +31 503617130.
E-mail addresses: firstname.lastname@example.org (M.R. van der Sijde),
email@example.com (A. Ng), firstname.lastname@example.org (J. Fu).
1University of Groningen, University Medical Centre Groningen, Department of
Genetics, PO Box 30.001, 9700 RB Groningen, The Netherlands. Tel.: +31 503617130.
2Gastrointestinal Unit, Gray Jackson 8, Room GRJ-825, Massachusetts General Hospital
and Harvard Medical School, 55 Fruit Street, Boston, MA 02114, USA. Tel.: +1 617 724
0925-4439/© 2014 Elsevier B.V. All rights reserved.
Contents lists available at ScienceDirect
Biochimica et Biophysica Acta
journal homepage: www.elsevier.com/locate/bbadis
have been developed to combine GWAS findings with GO terms [6,7],
protein–protein interactions , information from pathway databases,
such as the Kyoto Encyclopaedia of Genes and Genomes (KEGG) [9,
10], or co-occurrence/co-citation in the literature . Although these
methods have proven successful in prioritizing candidate genes and
identifying overrepresented pathways, they tend to be heavily biased
by the extent of prior knowledge available (including functional anno-
tation and curated information). Many disease-associated candidate
genes that lack prior biological pathway annotation, will unintentional-
ly ‘miss the mark’ using these annotation-based enrichment methods.
The past few decades have witnessed an enormous growth in the
quantity, quality, diversity and richness of human molecular and func-
tional data being generated. The advances in high-throughput technol-
metabolome, epigenome and microbiome across multiple disease
states, phenotypes, perturbation conditions and/or cell types. These
~omics profiling approaches can be used not only to evaluate and mon-
itor the change of individual biomolecules in diseases, but also to inves-
tigate the genetic basis of inter-individual variation for the molecular
traits. Systems genetics, which evaluates this genetic basis, aims to re-
veal the genetic flow from DNA to the phenotype through intermediate
molecular traits. This strategy has been proposed as a powerful method
diseases [12,13]. To do so, the experimental design and analysis frame-
work can include multiple steps (Fig. 1): 1) association studies or link-
age analysis to identify genetic loci that underlie complex traits and
diseases; 2) genomics analysis to profile ~omics data and identify the
biomolecules that are of relevance; 3) genetical genomics studies to in-
vestigate the effect of genetic variants on multiple intermediate
molecular phenotypes and to illustrate the genetic variation of molecu-
lar traits; and 4) network modelling and causal inference analysis to
construct the molecular circuitry from genotype to phenotype. In this
review, we describe how systems genetics can help close the gap be-
tween genotype and phenotype. We focus on the genetic variation of
molecular traits and their dynamics (step 3) and on network modelling
andcausalinferenceanalysis (step4)toillustrate howsystemsgenetics
needed for effective diagnosis and development of therapeutic
2. Genetic variation of molecular traits
Genetic loci can enforce their risk via intermediate molecular traits,
such as gene expression, proteins and metabolites, resulting in inter-
individualvariationofbiomolecules thatisundergenetic control.Amo-
lecular trait can therefore be viewed as a quantitative trait and used in
yses, the genomic locus underlying the variation in the measured trait,
called quantitative trait locus (QTL), can be identified . The first
genome-wide genetic analysis on gene expression was performed in
haploid yeast segregants  and this proof-of-concept analysis dem-
onstrated a widespread genetic effect on gene expression. Subsequent
[16–20], and on other molecular levels, such as proteins [21–23], me-
tabolites [24–26] and methylation [27,28]. These studies have greatly
increased our knowledge of the functional consequences of genetic
Fig. 1. The systems genetics approach. Systems genetics combines genetical genomics, genetic association and genomics analyses to construct the causal inference from genotype to phe-
notype.Theapproachintegrates genotype information (g)with large-scale~omicsprofilingdata(o)andphenotype data (p)and can beusedto builda networkand infer causality forthe
M.R. van der Sijde et al. / Biochimica et Biophysica Acta 1842 (2014) 1903–1909
2.1. Expression QTL analysis
As transcripts are the first direct products of DNA, defining the ge-
netic variation on gene expression can provide crucial functional infor-
mation on genetic variation. The genomic loci that underlie variation in
gene expression are called expression quantitative trait loci (eQTL).
Based on the physical distance between the eQTL and the affected
gene, theeQTL can be classified ascis-eQTLor trans-eQTL.A cis-eQTL re-
within a distance of 250 kb to 1 Mb in natural populations and 1–5 Mb
in segregating populations) [29,30], while trans-eQTLs are SNP-gene
pairs that are further away from each other or which may even lie on
different chromosomes. eQTL analysis can have different implications
for interpreting GWAS associations and constructing networks, for in-
stance cis-eQTLs can help to prioritize causal variants and candidate
genes at disease-associated loci. Over 40% of disease-associated SNPs
are observed to have a cis-acting effect on gene expression [31,32].
Based on this reasoning, the cis-eQTL has been successfully used to se-
lect the novel but weak associations that do not pass the genome-
the sample size [33,34]. Another important potential advantage of
studying eQTLs is that they can provide insight into the disease mecha-
nism and underlying pathways. Trans-eQTLs can affect downstream
disease genes which were not identified by GWAS before and are be-
comingincreasingly important to resolve the molecular pathways lead-
ing to disease . However, a systematic identification of trans-eQTLs
is challenging, as a trans-eQTL is believed to be an indirect association
ical systems and the theory of phenotypic buffering, molecular factors
located further downstream in the pathway tend to have a smaller ef-
fect than the factors upstream . Recent advances in the develop-
ment of statistical frameworks for trans-eQTL analysis  and the
dramatically increased sample size  now allow for a systematic
identification of trans-eQTLs, providing insight into the downstream ef-
in 8086 human individuals has reported trans-effects for 233 disease-
associatedSNPs . Onestrikingexample is a SNPassociated with sys-
temic lupus erythematosus (SLE). This SNP was found not only to affect
the expression of the transcription factor IKZF1 (IKAROS family zinc
finger) in cis, but it also affects the expression of 10 different genes in
trans, including four genes involved in the complement system and six
in type 1 interferon response. Both of these processes are important
for SLE. This study demonstrated the power of trans-eQTL analysis in
identifying key regulators of disease and their downstream effects.
2.2. Genetic effects can propagate to different molecular levels
Expression quantitative trait locus studies have provento bepower-
ful in functional genomics. But in order to gain a full picture of the dis-
ease process, it is important to study how genetic variation propagates
from DNA to transcripts and further to other ~omics levels, such as pro-
teins and metabolites. The success of eQTL studies suggests that there is
a potential value in applying the QTL approach to other molecular traits
aswell.Forexample,Melzeretal.evaluated therole ofgenetic variation
on thelevels of 42 proteinsmeasured in 1200 individuals .Theyde-
tected cis-effects for eight proteins and trans-effects for one protein. Six
of these eightproteinscorrelated with inflammatoryormetabolic path-
ways and provided a mechanistic view on the disease process. On the
metabolism level, the most studied trait would be blood lipid levels
and up to date 157 loci have been robustly established [39,40]. These
loci account for ~10–20% of the variation in blood lipids and often un-
derlie susceptibility to many cardiovascular and metabolic traits. With
the developments in high-throughput technologies, like nuclear mag-
netic resonance (NMR) and mass spectrometry, an increasing number
of protein and metabolites can be quantified. This was demonstrated
by Karsten Suhre's group, who conducted a comprehensive and
systematic evaluation of genetic variance in blood metabolism. They
analysed over 250 metabolites in serum samples, thereby advancing
our knowledge of the molecular basis for many metabolic diseases
[25,41]. An example of a locus that was found to be associated with
blood metabolite concentrations is the FADS1 gene (fatty acid
desaturase 1). This locus is also associated with multiple complex dis-
eases and traits, including inflammatory bowel disease [42,43], heart
in this gene were observed to have a modest effect on the expression of
FADS1  as well as on blood HDL cholesterol, LDL cholesterol and tri-
glyceride levels . Their strongest effect was seen on phospholipids,
explaining up to 40% of the observed variation for this trait . This
findingsuggests thatinvestigating the genetic effecton multiple molec-
ular levels can provide a system-wide view on the downstream effects
of disease-associated SNPs and subsequent mechanistic insights into
the disease aetiology.
3. Dynamics of genetic variation
Genetic variants are observed to have a dynamic effect, depending
on the cell-type, tissue-type, developmental stage or environmental
condition [48,49]. Several studies have estimated the proportion of cis-
regulated gene expression specific to certain cell-types [50–52] or tis-
sues [53–56]. Although most cis-eQTLs show a concordant association
For example, Ding et al. found a difference in cis-eQTLs of only 1–5%
when comparing psoriatic skin samples and healthy skin samples, but
a difference of 30% when they compared eQTLs between skin and
lymphoblastoid cell lines (LCLs) . Similarly, 27.8% of the cis-eQTL
was found to be tissue-specific when comparing blood, liver, adipose
tissues and muscle , while 29% appeared to be tissue-specific in a
comparison study between LCLs, skin and fat tissue . These discov-
eries have huge implications for disease studies, as they emphasize
the importance of using gene expression data from tissues relevant for
the disease under study. For instance, for type 2 diabetes (T2D), eQTL
studies have mainly been focusing on liver, adipose tissue, muscle and
pancreatic B cells, and it was shown that T2D-associated SNPs were
enriched for cis-eQTLs in liver and adipose tissue . Not only cis-
eQTLs but also trans-eQTLs can exert cell-type or tissue-type specific ef-
fects. In a more recent study, T2D-associated SNPs were functionally
characterized through trans-eQTL analysis in five different tissue types
. The authors found an enrichment of a tissue-specific trans-effect
for T2D SNPs. This suggested the downstream effect of T2D-associated
genetic variants and the underlying pathways that may be active in
the disease pathogenesis.
3.1. The role of the environment
Environmental factors also play animportant role in shapingthe ge-
netic effect. Cadwell et al. [58,59] demonstrated that the interaction of
genes and environmentcan determine disease phenotypes in the intes-
tine. In a mouse model they found that the combination of an environ-
mental trigger (a norovirus infection) and a genetic variant in the
ATG16L1 gene(akeygeneinvolved in autophagy and previously associ-
ated with Crohn's disease) is required to generate thePaneth cell secre-
tory abnormalities characteristic of Crohn's disease in human. There is
also increasing evidence to indicate the important interplay between
tory bowel disease (IBD) and risk alleles in the IBD susceptibility genes
NOD2 and ATG16L1 show altered intestinal microbiota compositions,
with significant shifts in the frequencies of the Faecalibacterium
and Escherichia taxa . Host genetics may therefore substantially
influence the structure and establishment of the gut microbiome .
Understanding the complex interactions between genetics and en-
vironmental factors would greatly benefit from a systems genetics ap-
proach. An example of this is a study by Parks et al., who examined
M.R. van der Sijde et al. / Biochimica et Biophysica Acta 1842 (2014) 1903–1909
proach, they combined GWAS and eQTL analyses with the systematic
profiling of obesity traits and gut microbiota composition to assess
gene–environment interactions, which would not have been possible
with classic linkage studies or GWAS alone. Their findings showed
that host genetics have a profound effect on shaping the plasticity of
the gut microbiota in response to an environmental trigger. These ex-
amples thus illustrate the importanceandpromise of integratinggenet-
ic information with data from diverse multi-omics platforms towards
delineating and understanding disease biology.
3.2. Longitudinal studies
In order to determine the pathways and networks that are activated
during the onset and the development of diseases or during ageing, it
can be valuable to monitor biomolecular levels over time. Current
~omics profiling efforts usually capture a snapshot of a set of molecules
and their activity at defined time points. In order to establish the causal
effects and their downstream events, it is important to also explore and
take into account the time-delay or phase lag in the analysis of disease
hensive and very promising longitudinal multi-omics studies has been
the iPOP (integrative personal omics profiling) project led by Michael
Snyder . In this study, data was integrated from genomic,
transcriptomic, proteomic, metabolomic and autoantibody profiles
sampled from a single healthy individual over a 14-month period.
First, the genome of the individual was sequenced and SNPs, indels
and structural variants were determined. Based on these outcomes
theyassessed thegenetic diseaserisk, revealinganelevated riskfor cor-
onary artery disease, basal cell carcinoma, hypertriglyceridemia and
type 2 diabetes. Accordingly, next to the levels of transcripts, proteins
and metabolites, markers associated with high-risk disease phenotypes
striking observation was that an elevated glucose response onset was
tightly associated with respiratory syncytial virus (RSV) infection. Al-
though not yet proven, this observation has led to the interesting hy-
pothesis that viral infections could perhaps trigger an altered glucose
metabolic response that predisposes an individual to type 2 diabetes.
4. Network modelling and causal inference
It is clear that the underlying mechanisms of complex traits and dis-
eases have a complex basis, where the phenotype can be the result of
several genetic, molecular and environmental interactions. The next
step is to develop and apply user-friendly computational algorithms to
integrate and analyse the different phenotypic measurements and to
gain a clear picture of the biological complexity .
4.1. Network modelling
Network reconstruction is a powerful and widely used approach
that provides a flexible framework on which the complexity associated
with biological pathways can be built systematically. A biological net-
work depicts molecules (or a collection of molecules organized into
functional modules) in a given biological system as nodes and their in-
teractions as edges between the nodes. The edges can represent any
type of relationship or association, such as regulation, physical binding,
correlation or dependency between nodes. The reconstructed networks
can incorporate curated information as well as data-driven relation-
ships and can be inferred using different algorithms, including linear
models, Bayesian approaches and equation-based methods [66,67].
Network modellinghasbeen appliedtoa widerange of biological prob-
lems in the past few years, and has contributed to the discovery of sev-
eral disease genes and biomarkers [68–71]. For example, Bordbar et al.
used multi-omics data analysis (transcriptomics, proteomics and
metabolomics) to construct a genome-scale metabolic network, which
allowed them to determine metabolic modulators of macrophage acti-
vation . Similarly, Tannahill and colleagues generated a metabolic
map of lipopolysaccharide (LPS)-activated macrophages by combining
metabolomic and transcriptomic data. They found that glycolytic
genes were induced and were also correlated with expression profiles
of the altered metabolites . This led to the novel observation that
succinate is an important (but previously under-appreciated) metabo-
lite in innate immune signalling.
4.2. Inferring causality
When genetic information is incorporated into the network, it can
support the inference of causality for the interactions among different
traits. As genetic information flows from DNA to the intermediate phe-
notypes and eventually to the final phenotype (disease), there is a good
indication that thedisease-associated geneis also thecausalgenewhen
boththeintermediatephenotype andthedisease phenotypeare associ-
ated with the same genetic variant. For example, such a strategy helped
Musunuru and colleagues identify the SORT1 gene as the causal gene
underlying plasma LDL cholesterol and myocardial infarction . By
integrating eQTL data and lipid level measurements they found a
liver-specific eQTL effect on the expression of the SORT1 gene, suggest-
ing a transcriptional regulatory role of the disease-associated SNP. The
association of the same SNP with plasma LDL cholesterol also pointed
towards a role in lipid processes. Functional analyses indeed confirmed
the regulatory mechanism of modulating hepatic secretion of very low-
density lipoprotein (VLDL).
When a disease-associated SNP is also found to affect a molecular
trait, the causal relationship between the molecular phenotype and
the disease phenotype is not directly clear. Several models are possible
(Fig. 2): 1) the SNP can affect the molecular phenotype and the disease
the molecular phenotype, which causes the development of disease
(causal model); or 3) the SNP can cause the disease and in response to
the disease status the molecular phenotype is changed (reactive
model). Thus, to systematically construct the molecular pathway from
genotype to phenotype, mathematical models and computational algo-
rithms are needed to distinguish between the different models .
For example, Schadt et al. developed a likelihood-based causality
model selection (LCMS) method for this . In several succeeding
studies they demonstrated the use of their LCMS model to identify obe-
sity and atherosclerosis genes in mouse F2 populations [76–78].
Through the integration of DNA variation, gene expression and pheno-
typic information, the authors found that perturbations of predicted
obesity genes resulted in significant changes in obesity-related traits
. In the atherosclerosis study, the LCMS model was used with
mouse liver and adipose tissues, yielding hundreds of genes that tested
as causal for aortic lesions . One of these candidate causal genes,
C3ar1, was experimentally validated using a mouse knockout model in
the same study. Several causal genes were also found to be enriched
for human cardiovascular disease-related SNPs identified by the
Wellcome Trust Case Control Consortium . These studies clearly
models to identify causal relationships for complex diseases.
5. Challenges and future perspectives
Several studies have demonstrated thepower of systems geneticsin
understanding disease aetiology. However, there are still major chal-
lenges to face and several considerations need to be taken into account
when designing a systems genetics experiment.
The first challenge is the comprehensive acquisition of multi-
dimensional in-depth phenotype data. At the moment, for many dis-
eases, the data are scarce and heterogeneous in nature, making it diffi-
cult to study the dynamics of disease networks. Hence, the ideal
M.R. van der Sijde et al. / Biochimica et Biophysica Acta 1842 (2014) 1903–1909
systems geneticsexperiment should facilitate the generation of data for
the same individuals on multiple molecular levels and across different
conditions and tissue types. Obviously, the economics of generating
high-throughput data are an important factor here. Although the prices
for these technologies are dropping fast, they will need to decline fur-
ther in order to make longitudinal multi-omics studies readily afford-
able. It will be crucial for the years to come, to team up in consortia
and use and combine the data that is publicly available in biobanks
and databases. For instance, the Genotype-Tissue Expression (GTEx)
programme of the Broad Institute (www.broadinstitute.org/gtex/) 
aimstocreate a comprehensivepublic atlasof gene expression and reg-
ulation across multiple human tissues, and is a good first step in this di-
rection. Likewise, in recent years, various prospective cohort studies
have been set up, in which a group of individuals is followed over
time, eventually making it possible to predict potential disease out-
comes based on genetic risk, molecular biomarkers, physiological traits,
and environmental factors [81-82, 69]. As demonstrated by the work of
Michael Snyder's group, monitoring the development of potential dis-
eases over time, combined with ~omics profiling, can provide a better
understanding of the mechanism leading to disease .
The second challenge is related to the first and concerns the statisti-
cal power. In a systems genetics study, it is important to assess an ade-
quate number of samples to get enough power for association testing.
Although the sample size for GWAS and ~omics profilinghas greatly in-
the number of samples can still be much lower than the number of fac-
tors tested at the genome-wide level (e.g., genetic factors or molecular
factors). The result will be that, at the stringent significance level re-
quired to correct for multiple-testing, only the strong effects can be de-
tected and more modest effects will be missed. In order to detect the
more subtle effects, either the sample size needs to be increased dra-
matically or analysis approaches need to be applied to reduce the data
complexity. Network analysis can also be of help here, as it can be
or their participation in a common biological process or pathway. Rath-
er than examiningtheeffectof individual molecules, such modellingal-
lows one to examine the collective effect of groups of molecules,
providing increased power.
A third challenge is the complexity of the causal inference. As more
molecular phenotypes are measured, causative signals can be traced
eral of these phenotypic levels, then a proportion could be explained by
the same genetic variant. As mentioned earlier, various mathematical
models and algorithms can be used to characterize the relationships,
such as linear regression modelling or Mendelian randomization .
However, causal relationships are expected to be much more complex
than the models described in Fig. 2 . Many confounding factors re-
main unknown or undetectable. As a result, more advanced mathemat-
ical models and algorithms have been proposed, including Bayesian
networks  and structural equation modelling . New methods
need to be developed in the coming years that can take different as-
sumptions for different data-types into account.
For genomics data to contribute to our understanding of disease bi-
ology we must first take the hurdle of integrating all the information
that is now being generated. As we have discussed in this review, sys-
tems genetics has proven to be a promising approach to achieve this.
In this respect, it is crucial to shift from the classical gene-centred
view of disease biology to a network perspective, in which system-
wide interactionsinmultiplecelltypes,tissues, organs and theenviron-
ment together define the disease state. As systems genetics has the po-
tential to generate these network-wide views on disease aetiology, it is
likely to become a major tool in the development of personalized med-
icine and disease prevention in the near future.
M.S. and J.F. are supported by the Netherlands Organization for
Scientific Research (NWO-VENI grant 863.09.007 to J.F.) and the Sys-
tems Biology Centre for Metabolism and Ageing (SBC-EBA grant
853.00.110). A.N. is supported by the Massachusetts General Hospital
Department of Medicine investigator start-up funding.
We thank Jackie Senior for editing the manuscript.
 B.E. Stranger, E.A. Stahl, T. Raj, Progress and promise of genome-wide association
studies for human complex trait genetics, Genetics 187 (2011) 367–383.
 L.A. Hindorff, P. Sethupathy, H.A. Junkins, E.M. Ramos, J.P. Mehta, F.S. Collins, T.A.
Manolio, Potential etiologic and functional implications of genome-wide association
loci for human diseases and traits, Proc. Natl. Acad. Sci. U. S. A. 106 (2009) 9362–9367.
 H. Zhong, X. Yang, L.M. Kaplan, C. Molony, E.E. Schadt, Integrating pathway analysis
and genetics of gene expression for genome-wide association studies, Am. J. Hum.
Genet. 86 (2010) 581–591.
Fig. 2.Inferenceof causality.When two traits sharea geneticassociation,statisticalormathematical models can be usedtoinferthe relationship between the genotype (g),the molecular
trait, represented by ~omics profiling data (o), and the phenotype (p). An association can be reactive, causal or independent. The correlation between the residuals of the traits, after ac-
counting for genetic effects, can distinguish between the different causality models .
M.R. van der Sijde et al. / Biochimica et Biophysica Acta 1842 (2014) 1903–1909
 P. Jia, L. Wang, A.H. Fanous, C.N. Pato, T.L. Edwards, Z. Zhao, Network-assisted inves-
tigation ofcombined causalsignals from genome-wide association studies inschizo-
phrenia, PLoS Comput. Biol. 8 (2012) e1002587.
 Y. Lee, H. Li, J. Li, E. Rebman, I. Achour, K.E. Regan, E.R. Gamazon, J.L. Chen, X.H. Yang,
N.J. Cox, Y.A. Lussier, Network models of genome-wide association studies uncover
the topological centrality of protein interactions in complex diseases, J. Am. Med.
Inform. Assoc. 20 (2013) 619–629.
 M. Ashburner, C.A. Ball, J.A. Blake, D. Botstein, H. Butler, J.M. Cherry, A.P. Davis, K.
Dolinski, S.S. Dwight, J.T. Eppig, M.A. Harris, D.P. Hill, L. Issel-Tarver, A. Kasarskis, S.
Lewis, J.C. Matese, J.E. Richardson, M. Ringwald, G.M. Rubin, G. Sherlock, Gene
ontology: tool for the unification of biology The Gene Ontology Consortium, Nat.
Genet. 25 (2000) 25–29.
 T. Beissbarth, T.P. Speed, GOstat: find statistically overrepresented Gene Ontologies
within a group of genes, Bioinformatics 20 (2004) 1464–1465.
 E.J. Rossin, K. Lage, S. Raychaudhuri, R.J. Xavier, D. Tatar, Y. Benita, C. Cotsapas, M.J.
physically interact and suggest underlying biology, PLoS Genet. 7 (2011) e1001273.
 M. Kanehisa, S. Goto, KEGG: Kyoto encyclopedia of genes and genomes, Nucleic
Acids Res. 28 (2000) 27–30.
 C. Xie, X. Mao, J. Huang, Y. Ding, J. Wu, S. Dong, L. Kong, G. Gao, C.-Y. Li, L. Wei,
KOBAS 20: a web server for annotation and identification of enriched pathways
and diseases, Nucleic Acids Res. 39 (2011) W316–W322.
 S. Raychaudhuri, R.M. Plenge, E.J. Rossin, A.C.Y. Ng, S.M. Purcell, P. Sklar, E.M.
Scolnick, R.J. Xavier, D.Altshuler, M.J. Daly, Identifyingrelationshipsamonggenomic
disease regions: predicting genes at pathogenic SNP associations and rare deletions,
PLoS Genet. 5 (2009) e1000534.
 M. Civelek, A.J. Lusis, Systems genetics approaches to understand complex traits,
Nat. Rev. Genet. 15 (2014) 34–48.
 J.H. Nadeau, A.M. Dudley, Genetics Systems genetics, Science 331 (2011)
 R.C. Jansen, J.P. Nap, Genetical genomics: the added value from segregation, Trends
Genet. 17 (2001) 388–391.
 R.B. Brem, G. Yvert, R. Clinton, L. Kruglyak, Genetic dissection of transcriptional reg-
ulation in budding yeast, Science 296 (2002) 752–755.
 J.F. Ayroles, M.A. Carbone, E.A. Stone, K.W. Jordan, R.F. Lyman, M.M. Magwire, S.M.
Rollmann, L.H. Duncan, F. Lawrence, R.R.H. Anholt, T.F.C. Mackay, Systems genetics
of complex traits in Drosophila melanogaster, Nat. Genet. 41 (2009) 299–307.
 B.J. Bennett, C.R. Farber, L. Orozco, H.M. Kang, A. Ghazalpour, N. Siemers, M.
Neubauer, I. Neuhaus, R. Yordanova, B. Guan, A. Truong, W. Yang, A. He, P. Kayne,
P. Gargalovic, T. Kirchgessner, C. Pan, L.W. Castellani, E. Kostem, N. Furlotte, et al.,
A high-resolution association mapping panel for the dissection of complex traits
in mice, Genome Res. 20 (2010) 281–290.
 Y. Gilad, S. Rifkin, Revealingthe architecture of gene regulation: the promise of eQTL
studies, Trends Genet. 24 (2008) 408–415.
 S.B. Montgomery, E.T. Dermitzakis, From expression QTLs to personalized tran-
scriptomics, Nat. Rev. Genet. 12 (2011) 277–282.
 D.J. Gaffney, J.-B. Veyrieras, J.F. Degner, R. Pique-Regi, A.A. Pai, G.E. Crawford, M.
Stephens, Y. Gilad, J.K. Pritchard, Dissecting the regulatory architecture of gene ex-
pression QTLs, Genome Biol. 13 (2012) R7.
 D. Melzer, J.R.B. Perry, D. Hernandez, A.-M. Corsi, K. Stevens, I. Rafferty, F. Lauretani,
A. Murray, J.R. Gibbs, G. Paolisso, S. Rafiq, J. Simon-Sanchez, H. Lango, S. Scholz, M.N.
Weedon, S. Arepalli, N. Rice, N. Washecka, A. Hurst, A. Britton, et al., A genome-wide
association study identifies protein quantitative trait loci (pQTLs), PLoS Genet. 4
 A. Lourdusamy, S. Newhouse, K. Lunnon, P. Proitsi, J. Powell, A. Hodges, S.K. Nelson,
A. Stewart, S. Williams, I. Kloszewska, P. Mecocci, H. Soininen, M. Tsolaki, B. Vellas, S.
Lovestone, R. Dobson, Identification of cis-regulatory variation influencing protein
abundance levels in human plasma, Hum. Mol. Genet. 21 (2012) 3719–3726.
 L. Wu, S.I. Candille, Y. Choi, D. Xie, L. Jiang, J. Li-Pook-Than, H. Tang, M. Snyder,
Variation and genetic control of protein abundance in humans, Nature 499 (2013)
 C. Gieger, L. Geistlinger, E. Altmaier, M. Hrabé de Angelis, F. Kronenberg, T.
Meitinger, H.-W. Mewes, H.-E. Wichmann, K.M. Weinberger, J. Adamski, T. Illig, K.
Suhre, Genetics meets metabolomics: a genome-wide association study of metabo-
lite profiles in human serum, PLoS Genet. 4 (2008) e1000282.
 T. Illig, C. Gieger, G. Zhai, W. Römisch-Margl, R. Wang-Sattler, C. Prehn, E. Altmaier,
G. Kastenmüller, B.S. Kato, H.-W. Mewes, T. Meitinger, M.H. de Angelis, F.
Kronenberg, N. Soranzo, H.-E. Wichmann, T.D. Spector, J. Adamski, K. Suhre, A
genome-wide perspective of genetic variation in human metabolism, Nat. Genet.
42 (2010) 137–141.
 G. Nicholson, M. Rantalainen, J.V. Li, A.D. Maher, D. Malmodin, K.R. Ahmadi, J.H.
Faber, A. Barrett, J.L. Min, N.W. Rayner, H. Toft, M. Krestyaninova, J. Viksna, S.G.
Neogi, M.-E. Dumas, U. Sarkans, P. Donnelly, T. Illig, J. Adamski, K. Suhre, et al., A
genome-wide metabolic QTL analysis in Europeans implicates two loci shaped by
recent positive selection, PLoS Genet. 7 (2011) e1002270.
 J.R. Gibbs, M.P. van der Brug, D.G. Hernandez, B.J. Traynor, M.A. Nalls, S.-L. Lai, S.
Arepalli, A. Dillman, I.P. Rafferty, J. Troncoso, R. Johnson, H.R. Zielke, L. Ferrucci, D.L.
Longo, M.R. Cookson, A.B. Singleton, Abundant quantitative trait loci exist for DNA
methylation and gene expression in human brain, PLoS Genet. 6 (2010) e1000952.
 J.T. Bell, A. a Pai, J.K. Pickrell, D.J. Gaffney, R. Pique-Regi, J.F. Degner, Y. Gilad, J.K.
Pritchard, DNA methylation patterns associate with genetic and gene expression
variation in HapMap cell lines, Genome Biol. 12 (2011) R10.
 D.-J. de Koning, C.S. Haley, Genetical genomics in humans and model organisms,
Trends Genet. 21 (2005) 377–381.
 B. Ge, D.K. Pokholok, T. Kwan, E. Grundberg, L. Morcos, D.J. Verlaan, J. Le, V. Koka,
K.C.L. Lam, V. Gagné, J. Dias, R. Hoberman, A. Montpetit, M.-M. Joly, E.J. Harvey, D.
Sinnett, P. Beaulieu, R. Hamon, A. Graziani, K. Dewar, et al., Global patterns of cis
variation in human cells revealed by high-density allelic expression analysis, Nat.
Genet. 41 (2009) 1216–1222.
 D.L. Nicolae, E. Gamazon, W. Zhang, S. Duan, M.E. Dolan, N.J. Cox, Trait-associated
SNPs are more likely to be eQTLs: annotation to enhance discovery from GWAS,
PLoS Genet. 6 (2010) e1000888.
 A.C. Nica, S.B. Montgomery, A.S. Dimas, B.E. Stranger, C. Beazley, I. Barroso, E.T.
Dermitzakis, Candidate causal regulatory effects by integration of expression QTLs
with complex trait genetic associations, PLoS Genet. 6 (2010) e1000895.
 K. Fransen, M.C. Visschedijk, S. van Sommeren, J.Y. Fu, L. Franke, E.A.M. Festen, P.C.F.
Stokkers, A.A.van Bodegraven, J.B.A. Crusius, D.W. Hommes, P. Zanen, D.J.de Jong, C.
Wijmenga, C.C. van Diemen, R.K. Weersma, Analysis of SNPs with an effect on gene
expression identifies UBE2L3 and BCL3 as potential new risk genes for Crohn's dis-
ease, Hum. Mol. Genet. 19 (2010) 3482–3488.
 H. Zhong,J.Beaulaurier, P.Y. Lum, C. Molony, X.Yang,D.J. Macneil, D.T. Weingarth,B.
Zhang, D. Greenawalt, R. Dobrin, K. Hao, S. Woo, C. Fabre-Suver, S. Qian, M.R. Tota,
M.P. Keller, C.M. Kendziorski, B.S. Yandell, V. Castro, A.D. Attie, et al., Liver and adi-
pose expression associated SNPs are enriched for association to type 2 diabetes,
PLoS Genet. 6 (2010) e1000932.
 R.S.N. Fehrmann, R.C. Jansen, J.H. Veldink, H.-J. Westra, D. Arends, M.J. Bonder, J. Fu,
P. Deelen,H.J.M. Groen, A.Smolonska,R.K.Weersma,R.M.W. Hofstra, W. a Buurman,
S. Rensen, M.G.M. Wolfs, M. Platteel, A. Zhernakova, C.C. Elbers, E.M. Festen, G.
Trynka, et al., Trans-eQTLs reveal that independent genetic variants associated
with a complex phenotype converge on intermediate genes, with a major role for
the HLA, PLoS Genet. 7 (2011) e1002197.
Beale, R.C.H. de Vos, M. Dijkstra, R.A. Scheltema, F. Johannes, M. Koornneef, D.
Vreugdenhil, R. Breitling, R.C. Jansen, System-wide molecular evidence for pheno-
typic buffering in Arabidopsis, Nat. Genet. 41 (2009) 166–167.
 X.He, C.K. Fuller, Y. Song, Q. Meng, B. Zhang,X.Yang, H. Li, Sherlock: detecting gene-
disease associationsby matchingpatterns of expression QTL and GWAS, Am. J.Hum.
Genet. 92 (2013) 667–680.
 H.-J. Westra, M.J. Peters, T. Esko, H. Yaghootkar, C. Schurmann, J. Kettunen, M.W.
Christiansen, B.P. Fairfax, K. Schram, J.E. Powell, A. Zhernakova, D.V. Zhernakova, J.H.
Veldink, L.H. Van den Berg, J. Karjalainen, S. Withoff, A.G. Uitterlinden, A. Hofman, F.
Rivadeneira, P.A.C. 't Hoen, et al., Systematic identification of trans eQTLs as putative
drivers of known disease associations, Nat. Genet. 45 (2013) 1238–1243.
 T.M. Teslovich, K. Musunuru, A.V. Smith, A.C. Edmondson, I.M. Stylianou, M. Koseki,
J.P. Pirruccello, S. Ripatti, D.I. Chasman, C.J. Willer, C.T. Johansen, S.W. Fouchier, A.
Isaacs, G.M. Peloso, M. Barbalic, S.L. Ricketts, J.C. Bis, Y.S. Aulchenko, G.
Thorleifsson, M.F. Feitosa, et al., Biological, clinical and population relevance of 95
loci for blood lipids, Nature 466 (2010) 707–713.
 C.J. Willer, E.M. Schmidt, S. Sengupta, G.M. Peloso, S. Gustafsson, S. Kanoni, A. Ganna,
J. Chen, M.L. Buchkovich, S. Mora, J.S. Beckmann, J.L. Bragg-Gresham, H.-Y. Chang, A.
Demirkan, H.M. Den Hertog, R. Do, L.A. Donnelly, G.B. Ehret, T. Esko, M.F. Feitosa,
et al., Discovery and refinement of loci associated with lipid levels, Nat. Genet. 45
 K. Suhre, S.-Y. Shin, A.-K. Petersen, R.P. Mohney, D. Meredith, B. Wägele, E. Altmaier,
P. Deloukas, J. Erdmann, E. Grundberg, C.J. Hammond, M.H. de Angelis, G.
Kastenmüller, A. Köttgen, F. Kronenberg, M. Mangino, C. Meisinger, T. Meitinger,
H.-W. Mewes, M.V. Milburn, et al., Human metabolic individuality in biomedical
and pharmaceutical research, Nature 477 (2011) 54–60.
 A. Franke, T. Balschun, C. Sina, D. Ellinghaus, R. Häsler, G. Mayr, M. Albrecht, M.
Wittig, E. Buchert, S. Nikolaus, C. Gieger, H.E. Wichmann, J. Sventoraityte, L.
Kupcinskas, C.M. Onnie, M. Gazouli, N.P. Anagnou, D. Strachan, W.L. McArdle, C.G.
Mathew, et al., Genome-wide association study for ulcerative colitis identifies risk
loci at 7q22 and 22q13 (IL17REL), Nat. Genet. 42 (2010) 292–294.
 L. Jostins, S. Ripke, R.K. Weersma, R.H. Duerr, D.P. McGovern, K.Y. Hui, J.C. Lee, L.P.
Schumm, Y. Sharma, C.A. Anderson, J. Essers, M. Mitrovic, K. Ning, I. Cleynen, E.
interactions have shaped the genetic architecture of inflammatory bowel disease,
Nature 491 (2012) 119–124.
 M. den Hoed, M. Eijgelsheim, T. Esko, B.J.J.M. Brundel, D.S. Peal, D.M. Evans, I.M.
Nolte, A.V. Segrè, H. Holm, R.E. Handsaker, H.-J. Westra, T. Johnson, A. Isaacs, J.
Yang, A. Lundby, J.H. Zhao, Y.J. Kim, M.J. Go, P. Almgren, M. Bochud, et al., Identifica-
tion of heart rate-associated loci and their effects on cardiac conduction and rhythm
disorders, Nat. Genet. 45 (2013) 621–631.
 W. Xie, A.R. Wood, V. Lyssenko, M.N. Weedon, J.W. Knowles, S. Alkayyali, T.L.
Assimes, T. Quertermous, F. Abbasi, J. Paananen, H. Häring, T. Hansen, O. Pedersen,
U. Smith, M. Laakso, J.M. Dekker, J.J. Nolan, L. Groop, E. Ferrannini, K.-P. Adam,
et al., Genetic variants associated with glycine metabolism and their role in insulin
sensitivity and type 2 diabetes, Diabetes 62 (2013) 2141–2150.
 A.K.Manning,M.-F. Hivert, R.A.Scott, J.L.Grimsby, N.Bouatia-Naji, H. Chen, D.Rybin,
C.-T. Liu, L.F. Bielak, I. Prokopenko, N. Amin, D. Barnes, G. Cadby, J.-J. Hottenga, E.
Ingelsson, A.U. Jackson, T. Johnson, S. Kanoni, C. Ladenvall, V. Lagou, et al., A
genome-wide approach accounting for body mass index identifies genetic variants
influencing fasting glycemic traits and insulin resistance, Nat. Genet. 44 (2012)
 S. Kathiresan, O. Melander, C. Guiducci, A. Surti, N.P. Burtt, M.J. Rieder, G.M. Cooper,
C. Roos, B.F. Voight, A.S. Havulinna, B. Wahlstrand, T. Hedner, D. Corella, E.S. Tai, J.M.
Ordovas, G. Berglund, E. Vartiainen, P. Jousilahti, B. Hedblad, M.-R. Taskinen, et al.,
Six new loci associated with blood low-density lipoprotein cholesterol, high-
density lipoprotein cholesterol or triglycerides in humans, Nat. Genet. 40 (2008)
 M. Ackermann, W. Sikora-Wohlfeld, A. Beyer, Impact of natural genetic variation on
gene expression dynamics, PLoS Genet. 9 (2013) e1003514.
M.R. van der Sijde et al. / Biochimica et Biophysica Acta 1842 (2014) 1903–1909
 M. Francesconi, B. Lehner, The effects of genetic variation on gene expression dy- Download full-text
namics during development, Nature 505 (2014) 208–211.
 A.S. Dimas, S. Deutsch, B.E. Stranger, S.B. Montgomery, C. Borel, H. Attar-Cohen, C.
Ingle, C. Beazley, M. Gutierrez Arcelus, M. Sekowska, M. Gagnebin, J. Nisbett, P.
Deloukas, E.T. Dermitzakis, S.E. Antonarakis, Common regulatory variation impacts
gene expression in a cell type-dependent manner, Science 325 (2009) 1246–1250.
 A. Gerrits, Y. Li, B.M. Tesson, L.V. Bystrykh, E. Weersing, A. Ausema, B. Dontje, X.
Wang, R. Breitling, R.C. Jansen, G. de Haan, Expression quantitative trait loci are
highly sensitive to cellular differentiation state, PLoS Genet. 5 (2009) e1000692.
 J.E. Powell, A.K. Henders, A.F. McRae, M.J. Wright, N.G. Martin, E.T. Dermitzakis, G.W.
Montgomery, P.M. Visscher, Genetic control of gene expression in whole blood and
lymphoblastoid cell lines is largely independent, Genome Res. 22 (2012) 456–466.
 A.L. Price, A. Helgason, G. Thorleifsson, S.A. McCarroll, A. Kong, K. Stefansson, Single-
tissue and cross-tissue heritability of gene expression via identity-by-descent in re-
lated or unrelated individuals, PLoS Genet. 7 (2011) e1001317.
 A.C. Nica, L. Parts, D. Glass, J. Nisbet, A. Barrett, M. Sekowska, M. Travers, S. Potter, E.
Grundberg, K. Small, A.K. Hedman, V. Bataille, J. Tzenova Bell, G. Surdulescu, A.S.
Dimas, C. Ingle, F.O. Nestle, P. di Meglio, J.L. Min, A. Wilk, et al., The architecture of
gene regulatory variation across multiple human tissues: the MuTHER study, PLoS
Genet. 7 (2011) e1002003.
 J. Fu, M.G.M. Wolfs, P. Deelen, H.-J. Westra, R.S.N. Fehrmann, G.J. Te Meerman, W.A.
Buurman, S.S.M. Rensen, H.J.M. Groen, R.K. Weersma, L.H. van den Berg, J. Veldink,
R.A. Ophoff, H. Snieder, D. van Heel, R.C. Jansen, M.H. Hofker, C. Wijmenga, L.
variation of gene expression, PLoS Genet. 8 (2012) e1002431.
 J. Ding, J.E. Gudjonsson, L. Liang, P.E. Stuart, Y. Li, W. Chen, M. Weichenthal, E.
Ellinghaus, A. Franke, W. Cookson, R.P. Nair, J.T. Elder, G.R. Abecasis, Gene expres-
sion in skin and lymphoblastoid cells: Refined statistical method reveals extensive
overlap in cis-eQTL signals, Am. J. Hum. Genet. 87 (2010) 779–789.
 S.C. Elbein, E.R. Gamazon, S.K. Das, N. Rasouli, P.A. Kern, N.J. Cox, Genetic risk factors
for type 2 diabetes: a trans-regulatory genetic architecture? Am. J. Hum. Genet. 91
 K. Cadwell, J.Y. Liu, S.L. Brown, H. Miyoshi, J. Loh, J.K. Lennerz, C. Kishi, W. Kc, J.A.
Carrero, S. Hunt, C.D. Stone, E.M. Brunt, R.J. Xavier, B.P. Sleckman, E. Li, N.
Mizushima, T.S. Stappenbeck, H.W. Virgin, A key role for autophagy and the autoph-
agy gene Atg16l1 in mouse and human intestinal Paneth cells, Nature 456 (2008)
 K. Cadwell, K.K. Patel, N.S. Maloney, T.-C. Liu, A.C.Y. Ng, C.E. Storer, R.D. Head, R.
Xavier, T.S. Stappenbeck, H.W. Virgin, Virus-plus-susceptibility gene interaction de-
termines Crohn's disease gene Atg16L1 phenotypes in intestine, Cell 141 (2010)
 D.N. Frank, C.E. Robertson, C.M. Hamm, Z. Kpadeh, T. Zhang, H. Chen, W. Zhu, R.B.
Sartor, E.C. Boedeker, N. Harpaz, N.R. Pace, E. Li, Disease phenotype and genotype
are associated with shifts in intestinal-associated microbiota in inflammatory
bowel diseases, Inflamm. Bowel Dis. 17 (2011) 179–184.
 A.D. Kostic, R.J. Xavier, D. Gevers, The microbiome in inflammatory bowel diseases:
current status and the future ahead, Gastroenterology (2014), http://dx.doi.org/10.
 B.W. Parks, E. Nam, E. Org, E. Kostem, F. Norheim, S.T. Hui, C. Pan, M. Civelek, C.D.
Rau, B.J. Bennett, M. Mehrabian, L.K. Ursell, A. He, L.W. Castellani, B. Zinker, M.
Kirby, T.A. Drake, C.A. Drevon, R. Knight, P. Gargalovic, et al., Genetic control of obe-
sity and gut microbiota composition in response to high-fat, high-sucrose diet in
mice, Cell Metab. 17 (2013) 141–152.
 Z. Bar-Joseph, A. Gitter, I. Simon, Studying and modelling dynamic biological pro-
cesses using time-series gene expression data, Nat. Rev. Genet. 13 (2012) 552–564.
 R. Chen, G.I. Mias, J. Li-Pook-Than, L. Jiang, H.Y.K. Lam, R. Chen, E. Miriami, K.J.
Karczewski, M. Hariharan, F.E. Dewey, Y. Cheng, M.J. Clark, H. Im, L. Habegger, S.
Balasubramanian, M. O'Huallachain, J.T. Dudley, S. Hillenmeyer, R. Haraksingh, D.
Sharon, et al., Personal omics profiling reveals dynamic molecular and medical phe-
notypes, Cell 148 (2012) 1293–1307.
 A.R. Joyce, B.Ø. Palsson, The model organism as a system: integrating “omics” data
sets, Nat. Rev. Mol. Cell Biol. 7 (2006) 198–210.
 G. Karlebach, R. Shamir, Modelling and analysis of gene regulatory networks, Nat.
Rev. Mol. Cell Biol. 9 (2008) 770–780.
 D. Marbach, J.C. Costello, R. Küffner, N.M. Vega, R.J. Prill, D.M. Camacho, K.R. Allison,
M. Kellis, J.J. Collins, G. Stolovitzky, Wisdom of crowds for robust gene network in-
ference, Nat. Methods 9 (2012) 796–804.
 M. Inouye, J. Kettunen, P. Soininen, K. Silander, S. Ripatti, L.S. Kumpula, E.
Hämäläinen, P. Jousilahti, A.J. Kangas, S. Männistö, M.J. Savolainen, A. Jula, J.
Leiviskä, A.Palotie, V.Salomaa,M. Perola, M. Ala-Korpela,L. Peltonen,Metabonomic,
transcriptomic, and genomic variation of a population cohort, Mol. Syst. Biol. 6
 R.R. Nayak, M. Kearns, R.S. Spielman, V.G. Cheung, Coexpression network based on
natural variation in human gene expression reveals gene interactions and functions,
Genome Res. (2009) 1953–1962.
 A.J. Walley, P. Jacobson, M. Falchi, L. Bottolo, J.C. Andersson, E. Petretto, A.
Bonnefond, E. Vaillant, C. Lecoeur, V. Vatin, M. Jernas, D. Balding, M. Petteni, Y.S.
Park, T. Aitman, S. Richardson, L. Sjostrom, L.M.S. Carlsson, P. Froguel, Differential
coexpression analysis of obesity-associated networks in human subcutaneous adi-
pose tissue, Int. J. Obes. (Lond.) 36 (2012) 137–147.
 S.P. Smeekens, A. Ng, V. Kumar, M.D. Johnson, T.S. Plantinga, C. van Diemen, P.
Arts, E.T.P. Verwiel, M.S. Gresnigt, K. Fransen, S. van Sommeren, M. Oosting, S.-C.
Cheng, L.A.B. Joosten, A. Hoischen, B.-J. Kullberg, W.K. Scott, J.R. Perfect, J.W.M. van
as central for host defense against Candida albicans, Nat. Commun. 4 (2013) 1342.
 A. Bordbar, M.L. Mo, E.S. Nakayasu, A.C. Schrimpe-Rutledge, Y.-M. Kim, T.O.Metz,M.B.
Jones, B.C. Frank, R.D. Smith, S.N. Peterson, D.R. Hyduke, J.N. Adkins, B.O. Palsson,
Model-driven multi-omics data analysis elucidates metabolic immunomodulators of
macrophage activation, Mol. Syst. Biol. 8 (2012) 558.
 G.M. Tannahill, A.M. Curtis, J. Adamik, E.M. Palsson-McDermott, A.F. McGettrick, G.
Goel, C. Frezza, N.J. Bernard, B. Kelly, N.H. Foley, L. Zheng, A. Gardet, Z. Tong, S.S.
Jany, S.C. Corr, M. Haneklaus, B.E. Caffrey, K. Pierce, S. Walmsley, F.C. Beasley, et al.,
Succinate is an inflammatory signal that induces IL-1β through HIF-1α, Nature 496
 K. Musunuru, A. Strong, M. Frank-Kamenetsky, N.E. Lee, T. Ahfeldt, K.V. Sachs, X. Li,
H. Li, N. Kuperwasser, V.M. Ruda, J.P. Pirruccello, B. Muchmore, L. Prokunina-Olsson,
J.L. Hall, E.E. Schadt, C.R. Morales, S. Lund-Katz, M.C. Phillips, J. Wong, W. Cantley,
et al., From noncoding variant to phenotype via SORT1 at the 1p13 cholesterol
locus, Nature 466 (2010) 714–719.
 R.C. Jansen, B.M. Tesson, J. Fu, Y. Yang, L.M. McIntyre, Defining gene and QTL net-
works, Curr. Opin. Plant Biol. 12 (2009) 241–246.
 E.E. Schadt, J. Lamb, X. Yang, J. Zhu, S. Edwards, D. Guhathakurta, S.K. Sieberts, S.
Monks, M. Reitman, C. Zhang, P.Y. Lum, A. Leonardson, R. Thieringer, J.M. Metzger,
L. Yang,J.Castle, H. Zhu, S.F. Kash,T.A.Drake,A.Sachs, etal.,Anintegrativegenomics
approach to infer causal associations between gene expression and disease, Nat.
Genet. 37 (2005) 710–717.
 X. Yang, L. Peterson, R. Thieringer, J.L. Deignan, X. Wang, J. Zhu, S. Wang, H. Zhong, S.
Stepaniants, J.Beaulaurier, I.-M. Wang,R. Rosa, A.-M. Cumiskey, J.M.-J. Luo, Q. Luo, K.
Shah, J. Xiao, D. Nickle, A. Plump, E.E. Schadt, et al., Identification and validation of
genes affecting aortic lesions in mice, J. Clin. Invest. 120 (2010) 2414–2422.
 X. Yang, J.L. Deignan, H. Qi, J. Zhu, S. Qian, J. Zhong, G. Torosyan, S. Majid, B. Falkard,
R.R. Kleinhanz, J. Karlsson, L.W. Castellani, S. Mumick, K. Wang, T. Xie, M. Coon, C.
Zhang, D. Estrada-Smith, C.R. Farber, S.S. Wang, et al., Validation of candidate causal
genes for obesity that affect shared metabolic pathways and networks, Nat. Genet.
41 (2009) 415–423.
 Wellcome TrustCase Control Consortium, Genome-wideassociation study of 14,000
cases of seven common diseases and 3,000 shared controls, Nature 447 (2007)
 J. Lonsdale, J. Thomas, M. Salvatore, R. Phillips, E. Lo, S. Shad, R. Hasz, G. Walters, F.
Garcia, N. Young, B. Foster, M. Moser, E. Karasik, B. Gillard, K. Ramsey, S. Sullivan,
J. Bridge, H. Magazine, J. Syron, J. Fleming, et al., The Genotype-Tissue Expression
(GTEx) project, Nat. Genet. 45 (2013) 580–585.
 R.P. Stolk, J.G.M. Rosmalen, D.S. Postma, R.A. de Boer, G. Navis, J.P.J. Slaets, J. Ormel,
B.H.R. Wolffenbuttel, Universal risk factors for multifactorial diseases: LifeLines: a
three-generation population-based study, Eur. J. Epidemiol. 23 (2008) 67–74.
 A. Hofman, C.M. van Duijn, O.H. Franco, M.A. Ikram, H.L.A. Janssen, C.C.W. Klaver, E.J.
Kuipers, T.E.C. Nijsten, B.H.C. Stricker, H. Tiemeier, A.G. Uitterlinden, M.W. Vernooij,
J.C.M. Witteman, The Rotterdam Study: 2012 objectives and design update, Eur. J.
Epidemiol. 26 (2011) 657–686.
 G.D. Smith, S. Ebrahim, “Mendelian randomization”: can genetic epidemiology con-
tribute to understanding environmental determinants of disease? Int. J. Epidemiol.
32 (2003) 1–22.
 Y. Li, B.M. Tesson, G.A.Churchill, R.C. Jansen, Critical reasoningon causal inference in
genome-wide linkage and association studies, Trends Genet. 26 (2010) 493–498.
 E.E. Schadt, C. Molony, E. Chudin, K. Hao, X. Yang, P.Y. Lum, A. Kasarskis, B. Zhang, S.
Wang, C. Suver, J. Zhu, J. Millstein, S. Sieberts, J. Lamb, D. GuhaThakurta, J. Derry, J.D.
Storey, I. Avila-Campillo, M.J. Kruger, J.M. Johnson, et al., Mapping the genetic archi-
tecture of gene expression in human liver, PLoS Biol. 6 (2008) e107.
 S.V. Nuzhdin, M.L. Friesen, L.M. McIntyre, Genotype–phenotype mapping in a post-
GWAS world, Trends Genet. 28 (2012) 421–426.
M.R. van der Sijde et al. / Biochimica et Biophysica Acta 1842 (2014) 1903–1909