Each year, malaria kills about 1 million children and causes debilit-
ating illness in more than 500 million people1. Underlying this mas-
sive global health problem is a remarkable biological phenomenon,
the co-evolution of three eukaryotic genomes2–9 (Table 1). The disease
is caused by single-celled parasites of the genus Plasmodium, which
invade, and reproduce in, human erythrocytes. The parasites are then
transmitted from one person to another by blood-sucking mosquitoes
of the genus Anopheles.
The evolutionary ‘arms race’ between the parasite, its vector and the
human host is central to the problem of controlling disease. Plasmodium
populations are continually evolving to resist antimalarial drugs and
have sophisticated genetic mechanisms of evading the human immune
system, presenting a major problem for the development of a vaccine
against malaria2,7,10. Anopheles populations are likewise evolving to
resist the insecticides that are used to control malaria, but they also
have genetic defences against the parasite that might provide clues to
new control strategies11,12. Malaria has also been a strong force for recent
evolutionary selection in the human genome9,13, and uncovering all of
the human genetic factors that confer resistance to malaria would pro-
vide clues to the molecular basis of protective immunity that would be
invaluable for vaccine developers.
The genetic basis of human resistance to malaria can now be investi-
gated systematically at the level of the whole genome, by using genome-
wide association (GWA) analysis. In a typical GWA study, the genotype
of thousands of individuals is determined at the positions of half a mil-
lion or more single nucleotide polymorphisms (SNPs)14 (see page 728).
The ultimate goal of GWA analysis is to uncover all of the DNA sequence
variants that affect an individual’s risk of disease, without sequencing the
whole genome, by using statistical inferences based on common patterns
of variation in the genome.
An important question that can be addressed by GWA analysis is
why only some children develop severe malaria (that is, life-threatening
forms of the disease15,16) in communities in which every child is repeat-
edly infected with Plasmodium falciparum, the species of parasite that
is responsible for most deaths from malaria. Only a small proportion of
P. falciparum infections progress to severe malaria, and epidemiological
data indicate that about 25% of the risk is determined by human genetic
factors17 (Box 1). A typical study design is to recruit individuals with severe
malaria (cases) in a hospital setting and to recruit control individuals
essentially randomly from the general population. By comparing the fre-
quency of a set of SNPs in cases and controls, it is possible to estimate the
effect of different sequence variants on an individual’s risk of developing
severe malaria. Because the risk of developing severe malaria is probably
determined by many genetic factors and environmental factors operating
at different stages of infection, the effect of any one factor might be small,
so a large number of individuals must be studied to obtain statistically
Similar approaches could, in principle, be used to investigate the
emergence and molecular basis of drug resistance in Plasmodium popu-
lations or insecticide resistance in Anopheles populations. However, this
cannot be put into practice until genomic variation in Plasmodium and
Anopheles populations is better understood3,18–21. A complicating factor
for GWA studies of P. falciparum is that in a single infection, the para-
sites that are transmitted can have different genotypes, so an individual
who is infected frequently can carry a parasite population of great genetic
complexity. Another is that, in Africa, where malaria is most prevalent,
the P. falciparum genome has low levels of linkage disequilibrium19,21.
Linkage dis equilibrium is a fundamental concept to consider in GWA
analysis; it refers to the correlation between genotypes that is observed
at neighbouring positions in the genome. The lower the level of linkage
disequilibrium, the more positions in the genome need to be genotyped
for an effective GWA study. Recent technological advances in massively
paral lel sequencing of single DNA molecules22 might help to overcome
both of these problems, by enabling P. falciparum to be genotyped at a very
large number of pos itions in the genome and by helping to distinguish the
different parasite genotypes that can constitute a single infection.
In this Commentary, we describe how a global research network
has been established to investigate the effects of genomic variation
in humans on the biology and pathology of malaria. We focus on the
human genome because the tools for genotyping and the framework for
population genetics are further advanced than those for Plasmodium
and Anopheles species. More specifically, we outline the practical reas-
ons why malaria is more challenging to study by GWA analysis than
many other common diseases, and we describe how we have established
several projects that bring together large-scale studies carried out in
multiple locations to address key scientific questions. We also describe
the procedures that we use for standardizing and integrating data from
different investigators, as well as the policies that we have developed to
deal with issues of sample and data ownership, data release, intellectual
property and ethics.
Challenges of GWA studies of malaria
The genetic analysis of human resistance to malaria is challenging at
several levels, ranging from the practical and ethical issues of clinical
research in the developing world to the statistical genetic issues arising
from the great diversity of the populations that are affected.
A global network for investigating the
genomic epidemiology of malaria
The Malaria Genomic Epidemiology Network*
Large-scale studies of genomic variation could assist efforts to eliminate malaria. But there are scientific,
ethical and practical challenges to carrying out such studies in developing countries, where the burden of
disease is greatest. The Malaria Genomic Epidemiology Network (MalariaGEN) is now working to overcome
these obstacles, using a consortial approach that brings together researchers from 21 countries.
*A list of participants and affiliations appears in the online version of the paper at www.nature.com/nature.
NATURE|Vol 456|11 December 2008|doi:10.1038/nature07632
Recruiting a large number of individuals with severe malaria presents
challenges because most of the burden of malaria falls on poor com-
munities with underfunded health services and no systematic medical
records. A considerable proportion of children with severe malaria die
within hours of reaching a hospital; therefore, for the clinical phenotype of
malaria to be classified properly, research information must be gathered at
the time of hospital admission. This implies considerable responsibilities
on the part of the research team for ensuring standards of medical care,
particularly in a resource-poor setting. Also, it is not feas ible to take large
amounts of blood from children who are ill with malaria, many of whom
are anaemic, so it is often necessary to use whole-genome amplification to
obtain enough DNA for genotyping at numerous SNP positions. This can
reduce genotyping efficiency and thus diminish statistical power, making
an even larger sample size necessary23.
In addition, designing an appropriate ‘SNP genotyping’ strategy for
GWA studies of malaria is complicated by the large amount of genomic
variation in Africa. Because of the low levels of linkage disequilibrium
in populations in Africa, genotypes need to be sequenced at the pos-
itions of more SNPs than in studies of European populations. On the
basis of the initial data from the International HapMap Project (http://
www.hapmap.org), it was estimated that a GWA study of about 1.5 mil-
lion SNPs in an African population would be approximately equivalent
to a study of 0.6 million SNPs in a European population, in terms of the
ability to tag a high proportion of common sequence variants6. But it is
difficult to estimate how many SNPs will be required to tag all common
variants until resequencing studies have generated a comprehensive list
of common sequence variants in different African populations24.
Furthermore, the ethnic diversity of African populations presents
numerous statistical challenges for GWA studies. Many African com-
munities consist of several ethnic groups, and minor differences in the
ethnic composition of the case groups and the control groups can lead
to false-positive genetic associations. To exclude such artefacts, studies
need to be designed carefully, and statistical genetic methods that correct
for population structure need to be applied25.
Another problem arises from the genetic differences between popu-
lations in Africa, as opposed to within a single population. Signals of
association are not expected to be constant across GWA studies carried
out at different locations in Africa. For example, differences in haplotype
structure can result in variable signals of association around a causal
variant of a disease, particularly in genomic regions that have recently
undergone evolutionary selection. Also, different populations can har-
bour different factors that confer resistance to malaria. One example of
this is two resistance-associated forms of haemoglobin (haemoglobin S
and haemoglobin C) that result from different SNPs at adjacent locations
in HBB, the gene that encodes the β-chain of haemoglobin — these SNPs
have different patterns of distribution in West Africa26,27.
But such differences between populations can be also highly informa-
tive. For example, they can aid in uncovering genetic factors that have
evolved in specific populations and in investigating interactions between
genes and the environment. Importantly, differences in the patterns of
linkage disequilibrium between populations can help to distinguish a
causal variant from neighbouring polymorphisms. This is necessary
because many SNPs that have been associated with particular diseases
are not the causal variants but show an association signal simply as a
result of correlation with the causal variant (because of linkage disequi-
librium). Thus, GWA studies carried out at multiple sites in Africa could
provide a rich resource for identifying causal variants.
Developing a global research network
In the past, research into the human genetic factors that affect resist-
ance to malaria has been characterized by multiple research groups each
pursuing relatively small studies on their own samples. But the chance
of making a discovery, and replicating the finding, is greatly increased if
there are effective mechanisms for different research groups to share data
and thereby enlarge the number of samples that are studied. The concept
of forming a network for sharing data on the genomic epidemiology of
malaria — which was to become the Malaria Genomic Epidemiology
Network (MalariaGEN) — originated from work that was funded in 2003
by the Bill & Melinda Gates Foundation and by the UK Medical Research
Council. The purpose of this funding was to develop web-based software
that would allow the integration of clinical and genetic data collected by
different research groups. This funding also supported a workshop on
the ethical and ownership issues involved in sharing data, which was held
in Accra, Ghana, in January 2004 and attended by scientists and clinical
researchers from ten research groups in Africa.
Malaria GEN was established in 2005, with joint funding from the Bill
& Melinda Gates Foundation (through the Foundation for the National
Institutes of Health) and the Wellcome Trust, as part of the Grand Chal-
lenges in Global Health initiative28 (http://www.gcgh.org). The purpose
of this joint funding was to discover mechanisms of protective immunity
Table 1 | Malaria involves three eukaryotic genomes
Plasmodium falciparum* (ref. 2)
Haploid (transiently diploid in mosquitoes)
Plus the mitochondrial genome and
2.3 × 107
Anopheles gambiae† (ref. 3)
2 autosomal chromosomes,
X chromosome, Y chromosome
Plus the mitochondrial genome
2.8 × 108
22 autosomal chromosomes,
X chromosome, Y chromosome
Plus the mitochondrial genome
3.2 × 109
Genome size (base pairs)
Estimated number of
Databases of common
~105 unvalidated SNPs listed by dbSNP
~106 unvalidated SNPs listed by dbSNP ~6 × 106 validated SNPs listed by dbSNP
~3 × 106 SNPs investigated by the International
HapMap Project6 (http://www.hapmap.org)
~3 × 104 structural variants listed by the
Database of Genomic Variants
Resistance to severe malaria9:
mediated by mutations in HBB (resulting in
different forms of haemoglobin, HbS and HbC);
the presence of HBA1 or HBA2 (which encode
different forms of the α-chain of haemoglobin,
resulting in α+-thalassaemia);
by deficiency in G6PD
Resistance to antimalarials7:
to antifolates, mediated by DHFR and DHPS;
to chloroquine, mediated by CRT and MDR
Resistance to insecticides8:
to DDT and pyrethroids, mediated by the
kdr allele, a variant of a gene encoding a
voltage-gated sodium channel
CRT, chloroquine-resistance transporter gene; DDT, dichlorodiphenyltrichloroethane; DHFR, dihydrofolate reductase gene; DHPS, dihydropteroate synthase gene; G6PD, glucose-6-phosphate dehydrogenase
gene; HBB, β-globin gene (which encodes the β-chain of haemoglobin); kdr, knockdown resistance gene; MDR, multidrug-resistance gene; SNP, single nucleotide polymorphism. *Plasmodium is a genus of
protozoan. There are more than 100 species, of which 5 can infect humans: Plasmodium falciparum, Plasmodium vivax, Plasmodium malariae, Plasmodium ovale and some forms of Plasmodium knowlesi. P. falciparum
is the most severe form and is responsible for most deaths from malaria. †Anopheles is a genus of mosquito. There are about 400 species, of which about 40 transmit the Plasmodium parasites that cause human
malaria. Anopheles gambiae is a major vector of P. falciparum.
NATURE|Vol 456|11 December 2008
to malaria by combining analysis of human genome variation with large-
scale epidemiological studies in malaria-endemic regions. Five objectives
necessary for achieving this goal were identified: building a global net-
work for sharing data on the genomic epidemiology of malaria; collect-
ing DNA and clinical data from individuals with different phenotypes
of malaria; characterizing genetic variation in populations in malaria-
endemic regions; identifying genetic variants that provide protection
against severe malaria; and defining the immunological mechanisms by
which such genetic variants exert their protective effect.
The group of researchers who came together to tackle these objectives,
the MalariaGEN investigators, are mainly leaders of clinical, epidemio-
logical or immunological research projects in malaria-endemic areas, and
they contribute samples and data to the MalariaGEN programme. Other
MalariaGEN investigators contribute expertise and technical resources
related to high-throughput analysis of genomic variation, statistical genet-
ics or biomedical ethics. The host institutions of Malaria GEN investigators,
the Malaria GEN partner institutions, are located in 15 malaria-endemic
countries and 6 other countries (for additional information, see http://
www.malariagen.net/resource/1), and the institutions in malaria-endemic
countries have well-established study sites, where individuals are recruited
to participate in research. Most of these study sites are in sub-Saharan
Africa: in Burkina Faso, Cameroon, Gambia, Ghana, Kenya, Malawi, Mali,
Nigeria, Senegal, Sudan and Tanzania. There are also Malaria GEN study
sites in Papua New Guinea, Sri Lanka, Thailand and Vietnam.
To address the complexities involved in setting up such a global research
network, MalariaGEN investigators agreed, at an inaugural meeting in
Oxford, United Kingdom, in July 2005, to establish the network in four
stages. The first stage was to establish a set of principles and processes,
agreed by all investigators, to regulate a central resource of DNA sam-
ples and phenotypic data (Box 2 and see http://www.malariagen.net/
resource/1). More specifically, this involved standardizing scientific defi-
nitions and procedures, enabling partners to gain secure access to the
data resource via the Internet, and developing rules about data sharing,
intellectual property and appropriate consent.
The second stage was to define a core scientific programme of large-
scale experiments and statistical analysis, which would use data and exper-
tise from multiple investigators, and the results of which would belong
jointly to all of the investigators involved. Projects that are part of this
core programme are called Consortial Projects (Box 2). There are four
such projects so far, and each has a specific objective and a plan of action
(Table 2). After a Consortial Project has been defined, each investigator
decides whether he or she wishes to contribute to the project.
The third stage was to find ways of assisting investigators in malaria-
endemic countries to develop clinical and epidemiological studies that
would advance the core scientific programme. Investigators were invited
to submit funding proposals for projects at their study sites that would
contribute to Consortial Projects, using the research infrastructure of
the local partner institution and founded on the scientific interests and
expertise of the local investigators. Funding was allocated after prop-
osals had been reviewed by a group of investigators that represented the
network as a whole (with members from Cameroon, Gambia, Ghana,
Italy, Kenya, Malawi, Mali, Sri Lanka, Sudan, Tanzania and the United
Kingdom). This group evaluated both the scientific design and the feas-
ibility of the clinical and epidemiological studies proposed, taking into
account the infrastructure and expertise of the local partner institution
and study site.
The fourth stage was to strengthen the capacity to manage data, and
to carry out statistical and genetic analyses, at partner institutions in
malaria-endemic countries. A fellowship programme in data analysis
was established. After an open application process, a data fellow was
appointed at each partner institution. Most of the MalariaGEN data fel-
lows work on the team of a MalariaGEN investigator and have responsi-
bilities for managing the team’s data. All data fellows receive training and
support in data management, statistical genetics and computing skills.
This training is provided by a team of expert statisticians, geneticists
and computer programmers who work at the Malaria GEN Resource
Centre, which is based at two locations in the United Kingdom, at the
University of Oxford and at the Wellcome Trust Sanger Institute near
Cambridge. Members of the Malaria GEN resource centre organize regu-
lar data-analysis workshops, both in the United Kingdom and at partner
institutions in malaria-endemic countries. These workshops provide
structured teaching, together with an opportunity for data fellows to
share their experiences and to analyse their own data with hands-on
assistance from an expert.
Dealing with data
Sharing data is a simple concept but, when many investigators and part-
ner institutions are involved, it can be complex to put into practice.
There is the technical issue of how to amalgamate data from different
research groups. There needs to be transparency about the ownership
and permitted uses of the data and samples contributed by investig-
ators. Procedures need to be established for releasing data and, where
appropriate, for protecting intellectual property. This section outlines
how MalariaGEN has dealt with each of these areas.
Standardizing and integrating data
Standardizing and integrating data from multiple study sites is central
to MalariaGEN’s mission. As an example, Consortial Project 1 (Table 2),
which is the core project of MalariaGEN’s programme, depends on there
being a standardized clinical definition of severe malaria. Severe malaria
consists of several overlapping clinical syndromes, often referred to as
In a malaria-endemic region, there is a large variation in the clinical
severity of infections with Plasmodium falciparum. When a young child
becomes infected, he or she usually becomes ill with fever but eventually
recovers. A small proportion of infections progress to severe malaria;
that is, forms of the disease that have life-threatening complications,
such as profound anaemia or cerebral malaria15,16. After repeated
infections, older children and adults acquire clinical immunity to malaria,
meaning that they can tolerate infection without developing symptoms.
The figure illustrates the likelihood of progression from infection
with P. falciparum to death for a young child living in a malaria-endemic
region, showing frequencies that are representative for such a
child. In any given situation, the frequencies will depend greatly on
environmental factors, such as mosquito biting rates. Despite the
importance of environment, human genetic factors are estimated
to account for approximately 25% of the variation among African
children in the risk of developing severe malaria17. Known genetic factors
that confer resistance to malaria, such as the allele that encodes the
HbS form of haemoglobin (ref. 9) (Table 1), account for only a small
proportion of this variation, implying that many genetic factors involved
in resistance remain to be uncovered17.
Such genetic factors might operate at any stage of the progression
from receiving a bite from a malaria-carrying mosquito to dying as a
result of severe malaria. For example, individuals who carry the HbS-
encoding allele have a tenfold lower risk of severe malaria than those
who do not carry this allele, and the mechanism of protection seems
to be that HbS acts to suppress the number of parasites in the blood.
Box 1 | Progression of malarial disease in a malaria-endemic region
Typical frequency for a child in a
Infected 10 times per year
Parasitaemic 50% of each year
Fever episodes twice per year
Probability of 3% per year
Probability of 1% per year Death
Parasites in the blood
Bitten by infectious mosquito
NATURE|Vol 456|11 December 2008
subphenotypes: these include cerebral malaria (which is characterized
by coma), profound anaemia and respiratory distress. Some genetic
factors confer resistance generally to severe malaria, whereas others
might be specific for a subphenotype. The clinical definition of severe
malaria therefore depends on a combination of observations, some of
which (for example, respiratory distress) can be quantified less precisely
than others (for example, anaemia, through measuring haemoglobin
concentration), and there is ongoing research into how to minimize
the diagnostic error rate. After consulting Malaria GEN investigators
— and after a joint meeting with the Severe Malaria in African Chil-
dren network16, in Yaoundé, Cameroon, in November 2005 — a stand-
ardized case report form was agreed (see http://www.malariagen.net/
resource/1). This form is not intended to replace the case report forms
used by individual investigators but rather to provide a template for
extracting core information from different clinical data sets in a stan-
dardized manner, while giving investigators the freedom to collect data
in the way that is most appropriate to their own research.
In a large research network, there will be site-to-site variation in the
way in which clinical and epidemiological information is recorded and
stored at the local level, so investigators and data fellows are encouraged
to have an active role in data standardization and integration. This is
facilitated by web-based software developed specifically for this pur-
pose by the MalariaGEN resource centre. Investigators collect data using
the database format that is best supported at their institution, and they
periodically upload their data via a secure, password-protected inter-
face to a personalized section of the MalariaGEN website, which cannot
be accessed by others. Tools are provided for checking data integrity
and for transferring data into the database for the relevant Consortial
Project. The process of data transfer generally requires the investigator
to recode or transform certain variables in their own data set to match
the format of the project database, and the web-based software assists
and documents this process.
MalariaGEN investigators are also working on the standardization
of immunological assays as part of Consortial Project 2, which involves
investigating the genetic determinants of the immune response in dif-
ferent populations and environmental settings (Table 2). In the first
phase of this project, antibody measurements are being carried out at a
central reference laboratory to ensure that data from different study sites
can be directly compared. In the long term, the project seeks to develop
robust methods and standardized reagents that will enable reference
laboratories to be established at partner institutions.
Sharing data and establishing rules of ownership
Malaria GEN is a data-sharing community in which independent investi-
gators with different projects and research objectives contribute to a cen-
tral repository of DNA samples and a central database of core phenotypic
data for each Consortial Project. General principles of data sharing and
ownership were agreed at the inaugural meeting of MalariaGEN (Box 2
and see http://www.malariagen.net/resource/1). The major findings of
each Consortial Project will be published in scientific journals, with all
investigators who contributed to the project listed as authors. In addition,
investigators are encouraged to analyse the data that have been gener-
ated from their own samples, and to incorporate any additional clinical
or experimental data that they have for these samples; these analyses
are then permitted to be published independently of the findings of the
One of the most important considerations when building the database
for each Consortial Project was protecting the anonymity of research
participants. The MalariaGEN database contains no personal identi-
fiers and is not linked to databases at local study sites. However, one of
MalariaGEN’s key principles is that investigators should be able to ana-
lyse data generated from samples that they contributed and to amalga-
mate these data with locally held phenotypic data. A standard opera ting
procedure was therefore developed to ensure that the local databases
held by partner institutions that contain data generated by MalariaGEN
are designed and used according to appropriate ethical guidelines (see
Releasing data and protecting data as intellectual property
Because the scientific benefits of GWA studies are cumulative, the value
of a single study can be increased substantially if the data for individual
subjects are available to the wider scientific community, provided that
the identity of these individuals is securely protected14,29. Malaria GEN’s
policy on this topic was developed in consultation with all Malaria GEN
investigators and with ethics-review boards at several Malaria GEN part-
ner institutions (see http://www.malariagen.net/resource/1). In broad
terms, the data-release policy seeks to permit research that is consistent
with the nature of informed consent and the uses of the samples agreed
by the relevant ethics-review boards. A key concern that arose from
the consultation was to guard against the data being used in a way that
might lead to any form of ethnic stigmatization. Another concern was
to ensure that the timeline for data release is fair for investigators in
malaria-endemic countries who have contributed resources and data
to a project, because these investigators generally have less capacity for
analysing genetic data than researchers in rich countries. Balancing the
benefits of prompt data release with the need to protect the interests
of partner institutions, Malaria GEN’s current policy is to release GWA
data 9 months after contributing investigators have had access to the
complete data set. Data are placed in the European Genotype Archive
(http://www.ebi.ac.uk/ega) and are then made available on application
MalariaGEN investigators have agreed on a set of principles and
processes for sharing samples and data. An important step was to
define several Consortial Projects, each of which has a specific objective
and project plan (Table 2). Investigators can control how their samples
and data are used by MalariaGEN by specifying which Consortial
Projects they wish to contribute to.
• Consortial Project: a project that uses data and expertise from multiple
investigators, is carried out with core MalariaGEN funds, and is agreed
by the Project Management Committee, the funding bodies and all of
the investigators taking part in the project.
• Contributing investigator: an investigator who contributes data,
samples or expertise to a Consortial Project.
• Investigator’s own analysis: analysis carried out by a contributing
investigator on his or her own samples, either using data generated by
MalariaGEN or facilitated in another way by MalariaGEN.
• The ownership of physical samples and clinical data contributed to
Consortial Projects remains with the contributing investigator.
• The contributing investigator can request that DNA samples be
returned at any stage after the agreed experiments have taken place
and can use the samples for purposes other than those of MalariaGEN.
• Genotyping data generated by a Consortial Project is fully accessible
to the investigator who contributed the samples and may be used for
the investigator’s own analysis.
• The contributing investigator is responsible for ensuring that samples
are taken with the participants’ informed consent and for gaining local
ethical approval, with support from the MalariaGEN ethics team.
• The laboratories that process samples and analyse data are
responsible for the safety and integrity of the samples, and for
maintaining security and confidentiality of information.
• Authorship of publications by MalariaGEN will reflect the
contributions of all who have provided data and expertise, in
accordance with normal academic practice.
• Data at the level of the individual from human GWA studies will be
made available to the scientific community through an independent
• Intellectual-property protection will be sought only if it will expedite
the translation of a scientific discovery into affordable health benefits
for the populations that are most in need.
• For further details of MalariaGEN consortial policies, see http://
Box 2 | Key elements of MalariaGEN’s policy
NATURE|Vol 456|11 December 2008
to an independent data-access committee (as described on the Malari-
aGEN Data Access web page, http://www.malariagen.net/access). As
an additional check and balance, a working group is being established
to represent partner institutions and ethics-review boards in malaria-
endemic countries, and this group will be kept informed about applic-
ations for access to data and consulted about any proposed changes to
the data-release policy.
It was important for MalariaGEN to develop guidelines on the cir-
cumstances in which data should be protected as intellectual prop-
erty before publication, with careful consideration of arguments for
and against patenting discoveries30. On the one hand, if a scientific
discovery could lead to health benefits, then every effort should be
made to make these benefits available to those who need them most, a
process that could involve patenting the discovery. On the other hand,
there is an argument for releasing data as openly as possible when there
are no immediate applications for improving health and when open
access to the data could drive innovations that might lead to health
benefits. Arguably, for genomic epidemiology data, the prompt release
of scientific findings is, in general, the appropriate course of action,
but occasionally there might be discoveries that are exceptions to this.
Malaria GEN’s current policy is that intellectual-property protection
should be sought if all three of the following conditions are satisfied:
the discovery must be directly relevant to a medical application; it must
be probable that the intellectual property will be licensed for develop-
ment immediately; and the discovery must have been shown to require
intellectual-property protection as a stimulus for further development
(see http://www.malariagen.net/resource/1). In such cases, intellectual
property will be licensed to non-profit organizations if possible. And, if
financial benefits arise, then MalariaGEN will seek to ensure that these
benefits flow to the communities who participated in the research.
Engaging with ethical issues
A range of ethical and social issues arise in establishing a network to
share data between investigators in many countries. Ensuring ethical
standards for the conduct of clinical research in developing countries
raises many complex issues31. And the accumulation of detailed gen-
omic information about individuals is raising new questions for society
in general32. This combination of ethical and social issues needs to be
addressed appropriately 33. Malaria GEN has therefore established a team
with expertise in medical ethics, which works with investigators and
partner institutions to assess the ethical and social issues at dif ferent
study sites, with the aim of establishing best practices for the ethical
conduct of research carried out by MalariaGEN. This ethics team also
develops training materials for investigators and ethics-review boards
and has held workshops in Kenya, Mali, Thailand and Vietnam. To
support investigators in tackling specific ethical issues and to gain an
understanding of local practices, members of the team have also visited
study sites in Cameroon, Gambia, Ghana, Kenya, Malawi, Mali, Papua
New Guinea, Senegal and Sudan.
One of the most important aspects of this work is to find effective ways
of communicating with research participants34. For example, when a very
sick child is brought from her village to a busy government hospital,
and her parents are asked whether part of the diagnostic blood sample
can be used for a research project, it is often difficult to convey the dis-
tinction between medical diagnosis and medical research. Terms such
as ‘research’, ‘genetics’, ‘laboratory’ and ‘database’ might be meaningless
unless a concise and effective way is found of translating these concepts
into the local language (by using examples and metaphors drawn from
local experience), without creating anxiety by information overload. After
consulting investigators and ethics-review board members, Malaria GEN
has developed a template and guidelines for obtaining informed consent
from participants in genetic studies of resistance to malaria (see http://
www.malariagen.net/resource/1). To understand how guidelines can be
put into practice most effectively, the ethics team is also undertaking empir-
ical research on the process of gaining informed consent at different study
sites, with the objective of establishing best practice across Malaria GEN
study sites, while being sensitive to local culture and practices.
The ethics team is also working to develop models of consultation at
the community level that are appropriate for diverse cultural settings.
A sensitive issue for many communities is the potential abuse of genetic
data relating to ethnicity, which could result in stigmatization. Quali-
tative research is being carried out to understand the perspectives of
communities and other stakeholders on the collection and use of infor-
mation about ethnicity in genomic epidemiology projects. The aim is to
develop guidelines for the publication and release of data about ethnicity
that will provide the maximum scientific benefit while safeguarding the
interests of participants and their communities.
Many of the ethical and social challenges confronting MalariaGEN stem
from the diversity inherent in a large scientific enterprise with partners
in rich and poor countries that span multiple disciplines, from clinical
research and community-based research to state-of-the-art genomics and
bioinformatics. Often, partners need to agree on an appropriate balance
between standardization and shared practices on the one hand, and diver-
sity and sensitivity to local circumstances on the other hand. MalariaGEN’s
procedures for data integration and guidelines for informed consent are
examples of this process.
In September 2008, an ambitious plan for the elimination of malaria was
announced — the Global Malaria Action Plan (http://www.rbm.who.int/
gmap). This plan, which is supported by major international develop-
ment agencies and governments around the world, seeks to halve the
number of malaria cases worldwide by 2010 and to eliminate deaths from
malaria almost completely by 2015. But it cannot succeed without effec-
tive insecticides and antimalarial drugs. And even if the plan’s goals for
the next decade are achieved, the chance of controlling and elimin ating
malaria over the long term will be greatly increased if an effective vaccine
Table 2 | MalariaGEN Consortial Projects
1. Analysis of human genome variants
associated with resistance or
susceptibility to severe malaria
Burkina Faso, Cameroon, Gambia,
Ghana (Navrongo and Kumasi),
Kenya, Malawi, Mali, Nigeria,
Papua New Guinea, Tanzania
Burkina Faso, Kenya, Mali, Senegal,
Sri Lanka, Sudan and Tanzania
• Study sites have recruited >10,000 cases and >10,000 ethnically matched controls
for genetic association studies of severe malaria, with family trios being recruited at
four study sites.
• GWA studies of >10,000 individuals to be completed in 2009 for three study sites.
2. Analysis of human genetic factors
that determine antibody responses
• Standardized assays for antibodies specific for malaria antigens are being developed.
• Assays will be used in a range of community-based studies with various
epidemiological designs, to test associations between candidate genes and
• A DNA repository is being established to allow detailed studies of human genome
variation in malaria-endemic populations.
• Each site has recruited ≥90 individuals and has ethical approval for sequence data
to be publicly released.
• A data-sharing collaboration has been set up between investigators working on
genetic linkage studies of the intensity of malaria infection and other quantitative traits.
3. Analysis of human genome variation
in populations in which malaria is
Burkina Faso, Cameroon, Gambia,
Ghana, Kenya, Mali, Nigeria,
Papua New Guinea, Sudan,
Tanzania and Vietnam
4. Genetic linkage studies of quantitative
traits associated with malaria
Ghana, Senegal and Thailand
NATURE|Vol 456|11 December 2008
18. Mu, J. et al. Genome-wide variation and identification of vaccine targets in the Plasmodium
falciparum genome. Nature Genet. 39, 126–130 (2007).
19. Mu, J. et al. Recombination hotspots and population structure in Plasmodium falciparum.
PLoS Biol. 3, e335 (2005).
20. Jeffares, D. C. et al. Genome variation and evolution of the malaria parasite Plasmodium
falciparum. Nature Genet. 39, 120–125 (2007).
21. Volkman, S. K. et al. A genome-wide map of diversity in Plasmodium falciparum. Nature
Genet. 39, 113–119 (2007).
22. Hillier, L. W. et al. Whole-genome sequencing and variant discovery in C. elegans. Nature
Methods 5, 183–188 (2008).
23. Teo, Y. Y. et al. Whole genome-amplified DNA: insights and imputation. Nature Methods
5, 279–280 (2008).
24. Bhangale, T. R., Rieder, M. J. & Nickerson, D. A. Estimating coverage and power for genetic
association studies using near-complete variation data. Nature Genet. 40, 841–843
25. Price, A. L. et al. Principal components analysis corrects for stratification in genome-wide
association studies. Nature Genet. 38, 904–909 (2006).
26. Agarwal, A. et al. Hemoglobin C associated with protection from severe malaria in the
Dogon of Mali, a West African population with a low prevalence of hemoglobin S. Blood 96,
27. Modiano, D. et al. Haemoglobin C protects against clinical Plasmodium falciparum malaria.
Nature 414, 305–308 (2001).
28. Varmus, H. et al. Grand challenges in global health. Science 302, 398–399 (2003).
29. Manolio, T. A. et al. New models of collaboration in genome-wide association studies:
the Genetic Association Information Network. Nature Genet. 39, 1045–1051 (2007).
30. Chokshi, D. A., Parker, M. & Kwiatkowski, D. P. Data sharing and intellectual property in a
genomic epidemiology network: policies for large-scale research collaboration. Bull. World
Health Organ. 84, 382–387 (2006).
31. Nuffield Council on Bioethics. The Ethics of Research Related to Healthcare in Developing
Countries. <http:/ /www.nuffieldbioethics.org/go/ourwork/developingcountries/
publication_309.html> (Nuffield Council on Bioethics, 2002).
32. Lunshof, J. E., Chadwick, R., Vorhaus, D. B. & Church, G. M. From genetic privacy to open
consent. Nature Rev. Genet. 9, 406–411 (2008).
33. Chokshi, D. & Kwiatkowski, D. Ethical challenges of genomic epidemiology in developing
countries. Genomics Soc. Policy 1, 1–15 (2005).
34. Chokshi, D. A. et al. Valid consent for genomic epidemiology in developing countries.
PLoS Med. 4, e95 (2007).
Acknowledgements MalariaGEN’s primary funding is from the Wellcome Trust
(grant number 077383/Z/05/Z) and from the Bill & Melinda Gates Foundation,
through the Foundation for the National Institutes of Health (grant number 566)
as part of the Grand Challenges in Global Health initiative. Initial work on the
web-based software was funded by the Bill & Melinda Gates Foundation (grant
number 29015) and the UK Medical Research Council (grant number G0200454).
The Wellcome Trust (Sanger Institute core funding) and the Medical Research
Council (grant number G0600230) provide additional support for genotyping,
bioinformatics and analysis. We thank H. Pearson and colleagues at Bird & Bird for
pro bono advice on intellectual property. The MalariaGEN Resource Centre is part of
the European Union Network of Excellence on the Biology and Pathology of Malaria
Parasites. Individuals who helped to establish MalariaGEN are acknowledged online
Author Information Reprints and permissions information is available at
www.nature.com/reprints. The authors declare no competing financial interests.
Correspondence should be addressed to Dominic Kwiatkowski
The new science of genomic epidemiology could assist these efforts
to eliminate malaria, by providing more effective ways of monitoring
the emergence of parasite resistance to antimalarial drugs and of mos-
quito resistance to insecticides, and by providing new leads for malaria
vaccine development based on a better understanding of the natural
mechanisms of protective immunity.
If genomic epidemiology is to make a contribution in this way, there
need to be mechanisms in place to help researchers both in malaria-
endemic countries and worldwide to pool their resources. Research
groups in malaria-endemic countries need access to the technical
expertise and infrastructure for the large-scale analysis of genomic
variation. And research groups worldwide need to combine forces to
analyse the massive amounts of data being generated by these studies,
leading the way for important discoveries to be made. The MalariaGEN
community is endeavouring to learn how to build and maintain the
relationships, shared values and best practices that underpin this new
type of scientific collaboration.
1. Snow, R. W., Guerra, C. A., Noor, A. M., Myint, H. Y. & Hay, S. I. The global distribution of
clinical episodes of Plasmodium falciparum malaria. Nature 434, 214–217 (2005).
Gardner, M. J. et al. Genome sequence of the human malaria parasite Plasmodium
falciparum. Nature 419, 498–511 (2002).
Holt, R. A. et al. The genome sequence of the malaria mosquito Anopheles gambiae. Science
298, 129–149 (2002).
Lander, E. S. et al. Initial sequencing and analysis of the human genome. Nature 409,
Venter, J. C. et al. The sequence of the human genome. Science 291, 1304–1351 (2001).
The International HapMap Consortium. A haplotype map of the human genome. Nature
437, 1299–1320 (2005).
Su, X., Hayton, K. & Wellems, T. E. Genetic linkage and association analyses for trait
mapping in Plasmodium falciparum. Nature Rev. Genet. 8, 497–506 (2007).
Hemingway, J., Field, L. & Vontas, J. An overview of insecticide resistance. Science 298,
Kwiatkowski, D. P. How malaria has affected the human genome and what human genetics
can teach us about malaria. Am. J. Hum. Genet. 77, 171–190 (2005).
10. Wootton, J. C. et al. Genetic diversity and chloroquine selective sweeps in Plasmodium
falciparum. Nature 418, 320–323 (2002).
11. Ranson, H. et al. Evolution of supergene families associated with insecticide resistance.
Science 298, 179–181 (2002).
12. Riehle, M. M. et al. Natural malaria infection in Anopheles gambiae is regulated by a single
genomic control region. Science 312, 577–579 (2006).
13. Sabeti, P. C. et al. Positive natural selection in the human lineage. Science 312, 1614–1620
14. Wellcome Trust Case Control Consortium. Genome-wide association study of 14,000
cases of seven common diseases and 3,000 shared controls. Nature 447, 661–678 (2007).
15. Marsh, K. et al. Indicators of life-threatening malaria in African children. N. Engl. J. Med.
332, 1399–1404 (1995).
16. Taylor, T. et al. Standardized data collection for multi-center clinical studies of severe
malaria in African children: establishing the SMAC network. Trans. R. Soc. Trop. Med. Hyg.
100, 615–622 (2006).
17. Mackinnon, M. J., Mwangi, T. W., Snow, R. W., Marsh, K. & Williams, T. N. Heritability of
malaria in Africa. PLoS Med. 2, e340 (2005).
NATURE|Vol 456|11 December 2008
The Malaria Genomic Epidemiology Network Download full-text
Lead Investigators Eric Akum Achidi1, Tsiri Agbenyega2, Stephen Allen3,4,
Olukemi Amodu5, Kalifa Bojang6, David Conway6, Patrick Corran7, Panos Deloukas8,
Abdoulaye Djimde9, Amagana Dolo9, Ogobara Doumbo9, Chris Drakeley10,11,
Patrick Duffy12,13, Sarah Dunstan14, Jennifer Evans2,15, Jeremy Farrar14,
Deepika Fernando16, Tran Tinh Hien14, Rolf Horstmann15, Muntaser Ibrahim17,
Nadira Karunaweera16, Gilbert Kokwaro18, Kojo Koram19, Dominic Kwiatkowski8,20,
Martha Lemnge21, Julie Makani22, Kevin Marsh18, Pascal Michon3, David Modiano23,
Malcolm E. Molyneux24, Ivo Mueller3, Theonest Mutabingwa12, Michael Parker25,
Norbert Peshu18, Chris Plowe26,27, Odile Puijalon28, Jiannis Ragoussis20, John Reeder3,
Hugh Reyburn10,11, Eleanor Riley10, Jane Rogers8, Anavaj Sakuntabhai28, Pratap
Singhasivanon29, Sodiomon Sirima30, Giorgio Sirugo6, Adama Tall31, Terrie Taylor26,32,
Mahamadou Thera9, Marita Troye-Blomberg33, Tom Williams18 & Michael Wilson19
Data Fellows Lucas Amenga-Etego19,34, Tobias O. Apinjoh1, Edith Bougouma30,
Rajika Dewasurendra16, Mahamadou Diakite9, Anthony Enimil2, Ayman Hussein17,
Deus Ishengoma21, Muminatou Jallow6, Enmoore Lin3, Alioune Ly31,
Valentina D. Mangano20,23, Alphaxard Manjurano10,11, Laurens Manning3, Carolyne
M. Ndila18, Vysaul Nyirongo24, Tom Oluoch18, Nguyen T. N. Quyen14, Prapat Suriyaphol35
& Ousman Toure9
Resource Centre Kirk A. Rockett (Lab Projects Lead)20, Aaron Vanderwal (Informatics
Lead)20, Taane Clark (Statistics Lead)8,20, Michael Parker(Ethics Lead)20,25, Rebecca
Wrigley (Network Development Lead)20, Dominic Kwiatkowski8,20 (Director),
Daniel Alcock8, Sarah Auburn8, David Barnwell20, Susan Bull20,25, Susana Campino8,
Jantina deVries20,25, Abier Elzein17,20, Julie Evans20, Kathryn Fitzpatrick20, Anita
Ghansah19,20, Angie Green20, Lee Hart20, Eliza Hilton20, Christina Hubbart20, Catherine
Hughes20, Anna E. Jeffreys20, Katja Kivinen8, Bronwyn MacInnis8, Magnus Manske8,
Gareth Maslen8, Marilyn McCreight20, Alieu Mendy20, Catherine Moyes20, Aceme
Nyika8, Claire Potter20, Paul Risley7, Kate Rowlands20, Miguel SanJoaquin20,24,
Kerrin Small20, Elilan Somaskantharajah8, Marryat Stevens20, YikYing Teo20 &
Project Management Committee Tsiri Agbenyega2, Dan Carucci36, Katharine Cook37,
Alan Doyle37, Ogobara Duombo9, Jeremy Farrar14, Michael Gottlieb36, Kevin Marsh18,
Odile Puijalon28, Terrie Taylor26,32 & Dominic Kwiatkowski (Chair)8,20
1The University of Buea, PO Box 63, Buea, South West Province, Cameroon. 2Kwame Nkrumah
University of Science and Technology, Private Mail Bag, Kumasi, Ghana. 3Papua New Guinea
Institute of Medical Research, PO Box 378, Madang, Papua New Guinea. 4Swansea Medical
School, Swansea University, Singleton Park, Swansea, West Glamorgan SA2 8PP, UK. 5Institute
of Child Health, College of Medicine, University of Ibadan, Ibadan, Nigeria. 6MRC Laboratories,
Atlantic Road, Fajara, PO Box 273, Banjul, Gambia. 7National Institute for Biological Standards and
Control, Blanche Lane, South Mimms, Potters Bar, Hertfordshire EN6 3QG, UK. 8The Wellcome
Trust Sanger Institute, Hinxton, Cambridge CB10 1SA, UK. 9The Malaria Research & Training
Centre, University of Bamako, PO Box 1805, Bamako, Mali. 10London School of Hygiene & Tropical
Medicine, Keppel Street, London WC1E 7HT, UK. 11Joint Malaria Programme, Kilimanjaro Christian
Medical Centre, PO Box 3010, Moshi, Tanzania. 12Genome Science Center, Sokoine University
of Agriculture, PO Box 3000, Chuo Kikuu, Morogoro, Tanzania. 13Seattle Biomedical Research
Institute, 307 Westlake Avenue North, Seattle, Washington 98109, USA. 14Oxford University
Clinical Research Unit, The Hospital for Tropical Diseases, 190 Ben Ham Tu, Quan 5, Ho Chi
Minh City, Vietnam. 15Department of Molecular Medicine, Bernhard Nocht Institute for Tropical
Medicine, Postfach 30 41 2, D-20324 Hamburg, Germany. 16Faculty of Medicine, University
of Colombo, PO Box 271, Kynsey Road, Colombo 8, Sri Lanka. 17Institute of Endemic Disease,
University of Khartoum, Medical Service Science Campus, PO Box 102, Khartoum, Sudan. 18Kenya
Medical Research Institute (KEMRI)–Wellcome Trust Programme, PO Box 230, Kilifi, Kenya.
19Noguchi Memorial Institute for Medical Research, University of Ghana, PO Box LG 581, Accra,
Ghana. 20Wellcome Trust Centre for Human Genetics, University of Oxford, Roosevelt Drive,
Oxford OX3 7BN, UK. 21National Institute for Medical Research, PO Box 9653, Dar es Salaam,
Tanzania. 22Muhimbili University of Health and Allied Sciences, PO Box 65001, Dar es Salaam,
Tanzania. 23University of Rome ‘La Sapienza’, Piazzale Aldo Moro 5, 00185 Rome, Italy. 24Malawi–
Liverpool–Wellcome Trust Clinical Research Programme, College of Medicine, University of
Malawi, PO Box 30096, Chichiri, Blantyre 3, Malawi. 25The Ethox Centre, Department of Public
Health and Primary Health Care, University of Oxford, Badenoch Building, Old Road Campus,
Headington, Oxford OX3 7LF, UK. 26Blantyre Malaria Project, PO Box 32256, Chichiri, Blantyre
3, Malawi. 27University of Maryland School of Medicine, 655 West Baltimore Street, Baltimore,
Maryland 21201, USA. 28Institut Pasteur, Unité d’Immunologie Moléculaire des Parasites, 28 Rue
du Dr Roux, 75724 Paris Cedex 15, France. 29Faculty of Tropical Medicine, Mahidol University,
420/6 Ratchawithi Road, Ratchathewi, Bangkok 10400, Thailand. 30Centre National de Recherche
et Formation sur le Paludisme, Avenue de l’Oubritenga, BP 2208, Ouagadougou 01, Burkina
Faso. 31lnstitut Pasteur de Dakar, BP 220 Dakar, Senegal. 32Michigan State University, Department
of Internal Medicine, College of Osteopathic Medicine, East Lansing, Michigan 48825, USA.
33The Wenner-Gren Institute, Stockholm University, SE-106 91 Stockholm, Sweden. 34Navrongo
Health Research Centre, PO Box 114, Navrongo, Ghana. 35Faculty of Medicine, Siriraj Hospital,
Mahidol University, 2 Prannok road, Siriraj, Bangkoknoi, Bangkok 10700, Thailand. 36Foundation
for the National Institutes of Health, 9650 Rockville Pike, Bethesda, Maryland 20814, USA.
37The Wellcome Trust, Gibbs Building, 215 Euston Road, London NW1 2BE, UK.