Poorer Nations Turn to Publicly Developed GM Crops

Abstract

Genetically modified crops are often framed as the products of multinational corporations, but in poorer nations it is public research that is vibrant and attempting their development.

Full text

NATURE BIOTECHNOLOGY VOLUME 23 NUMBER 1 JANUARY 2005 27
FEATURE
Poorer nations turn to publicly developed GM crops
Joel I Cohen
Genetically modified crops are often framed as the products of multinational corporations, but in poorer nations it is
public research that is vibrant and attempting their development.
The second conference on the Cartagena
Protocol on Biosafety will be held in May in
Montreal, Canada1. One goal of the confer-
ence will be to reconcile practical challenges
in implementing its articles concerning living
modified organisms around the globe, particu-
larly in developing nations. I present here the
findings of a study that was a joint effort of
partners from 15 developing countries on three
continents and the International Food Policy
Research Institute (IFPRI, Washington, DC,
USA) to analyze the current state of research,
regulation, genetic resources and institutional
roles in developing genetically modified (GM)
crops2. This study is meant to be representa-
tive of key trends, rather than comprehensive
in approach. Information from this type of
study, the first of its kind, will help scientists,
policy makers and regulators understand their
respective country’s public GM crop research
agenda, identify polices and regulatory needs
for specific GM events and provide a transpar-
ent picture of national research and regulation
for stakeholders. This effort in no way mini-
mizes the need for safety evaluation, but seeks
research and regulatory efficiencies and effec-
tiveness so that all benefit.
Diversity of transformed crops and
phenotypes
In the 15 countries studied (see Box 1 for
methodology), public research pipelines for
GM crops contained 201 genetic transforma-
tion events for 45 different crops. (An event
is defined as the stable transformation—the
incorporation of foreign DNA into a living
plant cell—undertaken by a single institute
among the participating countries, thereby
providing a unique crop-and-trait combina-
tion.) Data collection began in 2001; data
were evaluated in 2002, updated and finalized
through the end of 2003. These pipelines have
produced GM crops, including cereals, vegeta-
bles, root, tuber and oil crops, sugar and cotton.
Many are nearing or in confined trials; others
Joel I. Cohen is at the International Food Policy
Research Institute (IFPRI), Environment and
Production Technology Division, 2033 K Street,
NW, Washington DC, USA.
e-mail: j.cohen@cgiar.org
Box 1 Research approach
Given that the development of biotech products is knowledge- and resource-intensive, the
survey was directed to preselected national experts with unique expertise and knowledge
of biotech, biosafety and genetic resources owing to their positions and research. These in-
cluded senior research leaders in national agricultural institutes, universities and regulatory
organizations, external experts, biosafety specialists and decision makers. Unlike studies
that have sampled initiatives in all developing countries22, our survey was restricted to 15
countries—those that had advanced work on GM in the regulatory stage or had regulatory
procedures in place— allowing a thorough and comprehensive analysis of specific data.
The survey examined and verified peer-reviewed data collected from 15 countries and a
total of 62 research institutes for 13 criteria (Table 1): (i) country (ii) food and fiber crops
(iii) source of germ plasm (iv) gene group (v) gene (vi) phenotype category (vii) function
(viii) regulatory status (ix) regulatory status by year (x) lead research institutes (xi) collabo-
rating institutes (xii) institutional arrangement and (xiii) dissemination. This study focuses
on six types of data: first, the diversity of transformed crops and phenotypes; second, the
most important transgene groups; third, sources and types of genetic resources; fourth,
field safety and regulatory status; fifth, research collaboration; and sixth, advancement
and distribution of improved seeds. Crops were categorized and sorted following the United
Nations Food and Agriculture Organization (FAO, Rome) FAOSTAT crop classification23.
Information was collected for phenotypic trait expression, as categorized by the United
States Department of Agriculture Animal and Plant Health Inspection Service (APHIS,
Washington, DC, USA). The genetic resources used for transformation were analyzed to
determine whether public or private institutions developed these resources, and whether
their original material was local or foreign (imported). The full study2 included data from
Bulgaria, but these are not reported here because data from no other European country
were available for comparison.
To study the progress of GM crops through to commercialization, data were collected
by regulatory stage, emphasizing the most advanced events possible. Four stages were
used: experimental (transformation events that produce stable transgenic plants derived
from multiple generations at the laboratory/greenhouse/glasshouse scale); confined field
trials (transformation events expressing stable traits in small-scale, single or multilocation
confined trials); scale-up (transgenic plants advancing into larger, precommercial trials); or
commercial release (products marketed to farmers through privately or publicly owned seed
companies or other institutional mechanisms). For experimental stage entries, experts were
asked to identify only highly developed biotechnologies coming from laboratory, green-
house or glasshouse and to indicate in what stage of regulation their respective events were
most accurately placed.

28 VOLUME 23 NUMBER 1 JANUARY 2005 NATURE BIOTECHNOLOGY
FEATURE
are in later stages of field testing and seeking
broader approval.
Table 1 summarizes the data by country,
including total number of events, crop types
transformed and phenotypic category. (Eight
phenotypic categories were used: agronomic
properties, bacterial resistance, fungal resis-
tance, herbicide tolerance, insect resistance,
product quality, virus resistance and other.)
The percentage of different phenotypic
groups among the 201 transformation events
identified is presented in Figure 1. Over half of
the 201 transformation events involve single
genes that confer biotic resistance to either viral
or insect stresses to the host plant. In 11 events,
stacked genes (those that simultaneously con-
fer more than one trait) are being tested for
phenotypic combinations. Some countries are
working on five or fewer crops, whereas others,
such as China and South Africa, are working
on 15 or more.
The ten crops with the largest number of
transformation events are shown in Figure 2.
Although most transformation events have
focused on cereals, significant numbers of a
diverse range of transgenic vegetables, fruits,
roots and tubers have also been created.
Significant progress has also been achieved in
transforming orphan (noncereal food staples
and indigenous crops, including mung beans,
beans, chickpeas, cowpeas, lupin, cacao and
coffee). The greatest numbers of transforma-
tion events to date are for rice, potatoes, maize
and papaya. Cotton, which is used as an oil and
fiber crop, is shown for comparison with food
crops in Figure 2.
Geographical breakdown
The largest number of transformation events
were generated by the seven Asian countries
surveyed (109), followed by the four African
countries (54), and the four Latin American
countries (38). However, Brazil also reported
37 events contracted by the private sector work-
ing with Embrapa (Brasilia), a public research
institute associated with Brazil’s Ministry of
Agricultural (Brasilia), to address their mar-
ket needs. Asian countries have products in all
stages of the research pipeline, having made
significant commitments to GM crops3,4,
and are already achieving significant success
with insect-resistant GM cotton approvals (in
China and to a lesser degree in India, and lastly,
Indonesia). Despite the large number of trans-
formation events in development in Asia, only
the Philippines has approved a commercial feed
crop for production, and China allows cultiva-
tion and use of publicly developed transgenic
vegetables. Indonesia had approved commer-
Table 1 Transformation events grouped by country, crops and phenotypic category
Continent Countries No. eventsaCrops Phenotypic categoryb
Africa Egypt 17 Cotton, cucumber, maize, melons, potatoes, squash and marrow,
tomatoes, watermelons, wheat
AP, FR, FR/HT, HT, HT/IR, IR, OO, PQ, VR
Kenya 4Cotton, maize, sweet potatoes HT, HT/IR, OO, PQ, VR
South Africa 20 Apples, grapes, lupin, maize, melons, pearl millet, potatoes,
sorghum, soybeans, strawberry, sugar cane, tomatoes, indigenous
vegetables
AP, BR, FR, HT, HT/AP, IR, PQ, VR
Zimbabwe 5 Cotton, cowpeas, maize, sweet potatoes, tomatoes FR, HT/VR, VR
Asia China 30 Cabbage, chili, cotton, maize, melons, papayas, potatoes, rice,
soybeans, tomatoes
AP, FR, IR, VR
India 21 Cabbage, cauliflower, chickpeas, citrus, eggplant, mung beans,
muskmelon, mustard/rapeseed, potatoes, rice, tomatoes
AP, FR, HT/AP, IR, IR/BR, OO, PQ, VR
Indonesia 14 Cacao, cassava, chili pepper, coffee, groundnuts, maize, mung
beans, papayas, potatoes, rice, shallot, soybeans, sugar cane,
sweet potatoes
AP, FR, IR, PQ, VR
Malaysia 5Oil, palms, papayas, rice HT, IR, VR
Pakistan 5Cotton, rice HT, IR, PQ, VR
Philippines 17 Bananas and plantains, maize, mangoes, papayas, rice, tomatoes AP, OO, VR
Thailand 7 Cotton, papayas, pepper, rice AP, BR, IR, VR
Latin America Argentina 21 Alfalfa, citrus, potatoes, soybeans, strawberry, sunflowers, wheat AP, BR, FR, IR, IR/BR, OO, PQ, VR
Brazil 9Beans, maize, papayas, potatoes, soybeans AP, BR, FR, HT, IR, PQ, VR
Costa Rica 5Bananas and plantains, maize, rice AP, IR, VR
Mexico 3 Bananas and plantains, maize, potatoes IR, VR
Total 201
aAn event is defined as the stable transformation—incorporation of foreign DNA into a living plant cell—undertaken by a single institute among the participating countries, thereby
providing a unique crop and trait combination. bPhenotypes are defined as follows: AP, agronomic properties; BR, bacterial resistance; FR, fungal resistance; HT, herbicide tolerance;
IR, insect resistance; OO, other; PQ, product quality; VR, virus resistance.
Multiple
genes
4%
BR 3%
OO 5%
HT 5%
PQ 8%
FR 10%
AP 12%
IR 26%
VR 27%
Figure 1 Total events distributed by phenotype.
AP, agronomic properties; BR, bacterial
resistance; FR, fungal resistance; HT, herbicide
tolerance; IR, insect resistance; OO, Other; PQ,

NATURE BIOTECHNOLOGY VOLUME 23 NUMBER 1 JANUARY 2005 29
FEATURE
cial GM cotton, but it has now been taken off
the market.
Sub-Saharan Africa, with the exception
of South Africa, lacks many capabilities and
resources to advance such research5. Many
countries are just considering whether to
conduct research on, or to allow import of,
GM crops or products. Research capacity and
potential markets are evolving (e.g., for insect-
resistant cotton), albeit subject to uncertainties
regarding use and trade6 . Kenya and Egypt have
demonstrated competence in regulatory and
import approvals, but have still not approved
any crop for open testing or commercial use.
Phenotypic groupings
Table 2 presents five of the eight phenotypic
groups having the highest number of clearly
identified genes or gene groups. Where the
specific genes were not provided, the coun-
try’s description of the trait being developed
is retained.
On the basis of the study data, we identi-
fied three groups of genes that appear of suf-
ficiently robust utility and suitability for wide
use. The first gene group consists of Cry genes
from Bacillus thuringiensis (Bt) that confer
resistance to lepidopteran insects. The sec-
ond group consists of coat proteins of plant
viruses used for inducing virus resistance. And
the third consists of genes conferring herbicide
tolerance. Most other gene groups and their
associated phenotypic traits have not yet dem-
onstrated robust applicability in the field. For
example, no gene group has yet to confer effec-
tive fungal resistance, although much experi-
mental activity has been spent on investigating
the glucanases and chitinases.
Similarly, no group of genes has been
shown to reliably confer bacterial resistance in
the field, even though many investigators are
studying the effects of antimicrobial peptides.
Thus, success has been limited in developing
crops with traits other than insect resistance,
virus resistance and herbicide tolerance.
Among the genes and gene groups being
tested, the Cry genes, coat protein genes and
herbicide tolerance genes are most likely to
move through regulation with fewer require-
ments, assuming already packaged data are
accepted by the developing country in which
tests would occur. This is because numerous
safety reviews have been conducted on these
genes in several countries. However, this does
not rule out tests to address specific environ-
mental or biodiversity concerns, as such results
may not be transferable from one country to
another.
The more unusual genes shown in Table 2
include different types of insect-resistance
genes, replicase genes, antisense genes and
genes encoding antimicrobial peptides.
Most countries are focusing on genes that
are already available and have already been
Table 2 Genes and gene groups in five phenotypic categories
Phenotype category
Gene/gene group Number of eventsa
Insect resistance 51
Bt 35
Galanthus nivalis agglutinin (Snowdrop lectin) 5
Pin 4
Trypsin inhibitor 2
Bt and trypsin inhibitor 2
Gall midge resistance gene (Gm2) 1
Alpha amylase inhibitor 1
Not disclosed 1
Viral resistance 53
Coat protein 47
Replicase 3
Coat protein and reporter genes 1
Coat protein and replicase 1
Antisense to tomato yellow leaf curl virus 1
Fungal resistance 21
Glucanase, chitinase 6
Glucanase, PGIP2 2
Chitinase and ap24 antifungal protein 2
Chitinase 2
Blast resistance 2
Not disclosed 2
PGIP1 and PGIP2 isolated in South Africa 1
Grape resveratrol 1
Glucanase (PGIP3) 1
b32, PGIP2 and other selected antifungal genes 1
AP24, CH5b, GLN3 1
Herbicide tolerance 11
5-enolpyruvylshikimate-3-phosphate synthase (EPSPS) 2
BAR encoding phosphinothricin acetyltransferase 6
Acetohydroxyacid synthase (AHAS)2
PsbA encoding D1 polypeptide of photosystem II 1
Bacterial resistance 7
Xa21-resistance (R) gene 5
Unspecified antibacterial 1
Unspecified antimicrobial peptides 1
aAn event is defined as the stable transformation—incorporation of foreign DNA into a living plant cell—undertaken by a
single institute among the participating countries, thereby providing a unique crop and trait combination. PGIP, polyga-
lacturonase-inhibiting protein
0 102030405060
All otherOOHTPQFRAPIRVR
All other
35 crops
Sugar cane
Alfalfa
Wheat
Soybeans
Cotton
Tomatoes
Papayas
Maize
Potatoes
Rice
No. of events
Figure 2 Phenotype characteristics sorted by
number of transformation events among the top
10 crops in the study data set. AP, agronomic
properties; BR, bacterial resistance; FR, fungal
resistance; HT, herbicide tolerance; IR, insect
resistance; OO, Other; PQ, product quality; VR,
virus resistance.

30 VOLUME 23 NUMBER 1 JANUARY 2005 NATURE BIOTECHNOLOGY
FEATURE
characterized, but a few are also investing in
their own gene discovery and development,
such as South Africa, Malaysia, Brazil, India
and China.
Sources and types of genetic resources
Access to plant genetic resources that possess
acceptable agronomic performance and are
suitable for transformation is an important
influence on adoption of technology. For this
study, genetic resources constitute landraces,
varieties and finished lines produced or derived
from developing countries. Foreign resources
are those brought to a developing country by
an external entity. Public materials are those
from any form of public institution and private
materials are those from companies, as well as
commodity organizations operating for and
within specific developing countries.
Data from the study show that 85% of the
genetic resources used for transformation have
been derived locally from public materials (Fig.
3). Public genetic resources, defined as locally
adapted and well preferred by farmers, were
identified for 41 of 45 crops. Unlike private
materials, these genetic resources are usually
unencumbered by varietal or intellectual prop-
erty claims. The use and management of this
local material becomes, therefore, all the more
important.
Field safety, regulatory status and costs
When study data are explained on the basis of
four regulatory stages (experimental, confined
trials, scale up, commercial release; see Box 1
for explanation), a total of 127 transformation
events are at the experimental stage, 44 are in
confined trials, 22 in scale-up testing (mostly
in China) and 7 are in the commercial-release
stage (see Fig. 4).
Of the 44 events in confined testing, many
have been under examination for years, wait-
ing approval for scale-up or precommercial
trials. Part of the difficulty with advancement,
as agreed upon by study participants, is confu-
sion about the exact amount of data required
for a confined (risk management) versus open
(risk assessment) trial. When possible, larger
tests could be done in partnership with seed
companies or with government seed produc-
tion facilities to help share the regulatory costs
involved (see Box 2).
We do not know the number of initial trans-
formation events required to obtain the num-
ber of events in Figure 4. However, it is not
known whether 44 events in confined testing,
spread over many crops, traits and countries,
will be sufficient for selecting superior GM
material, to increase seed production and
satisfy food safety needs. Figure 5 shows the
breakdown of these events by phenotypes,
making the diversity of approaches very clear.
Implications of the numbers and phenotypes
finishing confined testing need further analy-
sis, as confirmation is possible following on the
2004 harvests and test results.
Research collaborations
Study participants collected information on the
type of collaboration developed (if any) and
plans for dissemination of research outputs.
Questions included the number of institutions
involved, the type of collaboration developed
and whether any plans exist for dissemination
of the GM seed or planting material. Some
research institutes sought partnerships to
complete development of GM research prod-
ucts and to move research through regulation
and onto public or private producers. Despite
expectations of benefits to the public sector,
few partnerships were developed, including
those with the private sector7.
On the basis of the data, partnerships appear
to be less common events (80 transformation
events, representing only 40% of the total).
Figure 3 Source of genetic resources.
Figure 4 Number of publicly derived
transformation events (in GM crops) classified by
regulatory stage and region.
Foreign private 1%
Foreign public 3%
Local private 4%
All others 7%
Local public 85%
0
10
20
30
40
50
60
70
80
Latin AmericaAsiaAfrica
Commercial
release
Scale-upConfined
field trial
Lab/greenhouse
No. of events
Box 2 Investing in development rather than research alone
To date, most investment in biotech has been made for research infrastructure,
collaboration and scientific capacity building without forseeing the need to provide for
meeting regulatory requirements, especially as biosafety funds would then be diverted
from research4. Study participants discussed the costs and regulatory requirements for
developing GM products with a view to establishing efficiencies, sharing information
and material so that public institutes can better comply with regulatory requirements
and better manage their costs.
At the Next Harvest conference held in The Hague, Netherlands, 2002, Maria Jose
Sampaio, Intellectual Property Secretariat for Embrapa, presented data from Brazil
on the cost of compliance for the regulatory approval of a single transformation event,
including initial greenhouse and field screening, field testing for environmental impact
and food safety. The cost of compliance per event varied from $700,000 in virus-
resistant papaya, to $4 million for herbicide-resistant soybeans. The higher cost per
event for herbicide-resistant soybeans is mainly due to the requirement for complete
animal studies2. Benjamin Odhiambo, plant pathologist at Kenya’s Agricultural
Research Institute (KARI), presented data for insect-resistant maize in Kenya. The cost
for completing initial regulatory information for the maize event is $160,000, of which
the major component is the cost of testing in contained structures2. However, these
figures are now being revised.
According to Ana Sittenfeld, senior scientist, at Costa Rica’s Center for Research in
Cell and Molecular Biology at the Univeristy of Costa Rica (San Jose), the cost for regu-
latory compliance (including field trials but not technology development and molecular
characterization) for virus-resistant rice in Costa Rica was $2.25 million2. These initial
estimates are for the state of knowledge and the current biosafety regulatory system in
the respective country at the time of the conference. To understand these matters more
fully, a more intensive research study is required.

NATURE BIOTECHNOLOGY VOLUME 23 NUMBER 1 JANUARY 2005 31
FEATURE
Single, public R&D institutions, working with-
out any form of collaboration, conducted the
largest proportion (60%) of research (Table 3).
Of the 80 transformation events created
through partnerships, the majority (48 events)
involved public-public collaboration, most
often between public research institutions in
the same country.
Public-private collaborations were respon-
sible for 21 GM crop events (10%), including a
number from African countries (Egypt, Kenya
and South Africa). The international private
sector is involved in the majority of these cases,
local seed companies playing a minor role.
Advancement and dissemination
Participating countries were asked to share
preliminary plans as to how GM crops will
be disseminated to farmers. Results from the
study indicate that in general, such plans have
not been established—44% of the scientists
indicated they do not yet have suitable seed
distribution mechanisms to reach farmers.
Another 23% said that they would rely on pub-
lic sector methods of dissemination, involving
the national agricultural research institutes, or
universities (Figure 6). Private sector partner-
ships were being contemplated for 7% of the
cases. Preliminary plans for advancement of
the remaining GM plants were not available
by the end of the study.
Lack of collaborative and partnership
arrangements reflect the paucity of options
available for the developing countries8. The
partnerships reported do not include time
needed for acceptance, to engage farmers from
early to final stages, and to meet appropriate
seed or plant material suppliers.
Quality of life and enhancing food
security
Public research included in this study targets
research that could enhance quality of life
in agricultural communities and includes
research on many basic food staples of impor-
tance to local economies. Some of the GM
crops reported could yield several quality of
life improvements (see Table 4):
• Reduction in the use of conventional pesti-
cides, which has quantifiable environmen-
tal and human health benefits, as well as a
reduction in application costs per acre. Of
the transformation events reported in our
study, 35 confer insect-resistant traits to
crops, reflecting the perceived importance
of pests on regional economies.
• Reduction in the use of other agrochemicals
widely used to fight virus, fungus or other
diseases. Eighty-four transformation events
target this area, which if brought to the
market successfully, should have an effect in
reducing costs and increasing production.
• Improved abiotic stress crop tolerance,
such as drought and salinity that place
limitations on poor farmers located in less
favored regions. Of the 201 events, 11 are
being developed in this promising area.
• Better product quality, such as prolonged
shelf life or enhanced product characteristics
(foods delivering alternative carbohydrate or
fat composition) that would improve trans-
portation and consumer appeal of crops.
Of 15 transformation events being devel-
oped for product qualities, 5 are in the area
of nutritional enhancement and 6 are to
prolong shelf life. The other 4 are for prod-
uct characteristics, such as increased sucrose.
There are also major public initiatives,
such as HarvestPlus, that seeks to reduce
micronutrient malnutrition to breed nutri-
ent-dense staple foods (http://www.harvest-
plus.org).
• Alternative and more efficient provision of
essential vitamins and vaccines. Nine trans-
formation events are being developed for
plant-based vaccine deployment.
Traits that increase crop yield would also
be expected to have spillover effects in local
economies through generation of direct and
indirect employment and increase in personal
income and food security. Many of the traits
and genes identified have this potential, espe-
cially those for insect resistance, virus resis-
tance, fungal resistance, herbicide tolerance,
bacterial resistance and agronomic properties.
However, this can be determined only in field
trials, as yield reduction can occur from the
introduction of genes, through either conven-
tional or GM technologies.
Crops and traits identified in this study indi-
cate the potential impact and importance of
transgenic products to agriculture in develop-
ing countries. In addition, as we know from
Table 3 Partnerships sorted by institutions and by total number of transformation
events created at each institution
Continent Country
Institutions Eventsa
Number With partners Total With partners
Africa Egypt 1 1 17 12
Kenya 1 1 4 4
Zimbabwe 4 3 5 3
South Africa 5 2 28 7
Asia Malaysia 2 1 5 3
Pakistan 3 3 5 5
Philippines 3 3 17 17
Thailand 3 2 7 6
Indonesia 6 2 24 5
China 9 1 30 1
India 14 121 1
Latin America Mexico 1 0 3 0
Brazil 2 2 9 7
Costa Rica 3 3 5 5
Argentina 4 3 21 4
Total 61 28 201 80
aAn event is defined as the stable transformation—incorporation of foreign DNA into a living plant cell—undertaken by a
single institute among the participating countries, thereby providing a unique crop and trait combination.
Figure 5 Phenotypic characterization of all
44 field trials. AP, agronomic properties; BR,
bacterial resistance; FR, fungal resistance; HT,
herbicide tolerance; IR, insect resistance; OO,
Other; PQ, product quality; VR, virus resistance.
BR 2%
FR/HT
2%
HT/VR
2%
OO 5%
PQ 5%
FR 7%
A
P 9%
HT 12%
IR 26%
VR 32%

32 VOLUME 23 NUMBER 1 JANUARY 2005 NATURE BIOTECHNOLOGY
FEATURE
their use in rich and developing countries,
many of the GM crop events reported here
have a history of food and environmental
safety9. Approximately 76% of the events hav-
ing a direct relation to quality-of-life traits,
establishing their importance to agricultural
development (Table 4).
Conclusions—getting to specifics
This study finds the public sector to be a com-
petent, but largely unproven, player for GM
crop production in developing countries.
Whether national policies in these countries
stimulate or deter research and technology for
publicly developed GM crops is unclear; the
official approval of a publicly reported trans-
formation event for insect-resistant cotton in
China appears an isolated occurrence.
All in all, this study surveyed GM crop
research conducted at 61 public research
institutes in 15 developing economies. These
institutes have demonstrated transformation
capabilities across 45 plants, within eight cat-
egories of different transgenic phenotypes, and
the ability to use such genes when transform-
ing local genetic resources.
As scientific capabilities and the number of
research institutes increase, so will the diver-
sity of crops and phenotypes. Greater attention
is needed, however, for specific events where
resources and knowledge are lacking to com-
plete efficacy and safety testing. Otherwise,
GM crops will remain in preliminary test-
ing. Indeed, on the basis of this study’s data,
we estimate that approximately 22% of the
201 transformation events created in public
research programs remain in confined testing
(Fig. 4).
In contrast to achievements in R&D, most
developing countries have only limited expe-
rience in compiling regulatory data; in fact, it
has become difficult to complete all regulatory
requirements. Although many research trends
in this report are positive, few transformed
crops have been released from confined into
precommercial testing or into use.
This can be attributed to several factors: first,
the overall isolation of public research insti-
tutes; second, the inability of public research
to meet food safety and environmental regu-
latory requirements and confusion regarding
regulatory standards between confined versus
open trials; third, lack of regional abilities
to exchange and evaluate regulatory data on
specific transgenes and crops; fourth, exper-
tise with public genetic resources but few
opportunities to use or evaluate proprietary
germ plasm; fifth, difficulties in planning for
advancement of specific products; sixth, lim-
ited progress in determining authorities and
frameworks for science-based decision mak-
ing; seventh, implementing processes arising
from the international level (e.g., the Cartagena
Protocol for Biosafety1,10) as well as at the
regional level (e.g., special needs confronting
Africa11); and eight, external political barriers
that either halt regulatory review (e.g., mora-
toriums in Thailand)12,13 or have implications
for world trade (e.g., impasse over GM crops
between the United States and Europe14,15).
Policy, research and regulatory options are
needed to expedite regulatory decisions and
testing of public GM crops15,16. The sooner
such evaluations occur, the faster GM crops
unsuitable for field application can be dis-
carded and successful GM crops moved for-
ward, thus saving public funds and minimizing
opportunity costs. This report facilitates mak-
ing specific recommendations by scientists,
policy makers, regulators and other stakehold-
ers striving to evaluate and foster development
of publicly derived GM plants.
Fully exploit genetic resources. Using agro-
nomically productive genetic resources for
transformation, and not just for ease of regen-
eration, will expedite public research. This
study reveals that access to proprietary genetic
resources in developing countries is extremely
limited; only 6% of all transformation events
used private material.
Does the high percentage of local trans-
formed material mean reliance or dependence
on public genetic resources or a deliberate
independence from protected varieties or com-
mercial germ plasm? This question is not easy
to answer, as both choices present benefits and
costs, and different opportunities to the research
institute. The ability to transform local, widely
used public or indigenous genetic resources
provides the potential for greater public and
farmer acceptance. Using high-performance
GM public germ plasm means that farmers will
not be prevented from saving seeds, nor will
they potentially be under monopoly pricing of
seeds. However, some private companies have
promised free rights to their genes in specific
crops, such as sweet potato and the rice genome
for public research.
Ensure research serves the public good.
Examination of potential benefits and genetic
All other 4%
Research discontinued 4%
Private sector
partnership 7%
IP audit dependent 8%
Biosafety and/or agronomic
research required 10%
Public sector mechanisms 23%
Not yet determined
44%
Table 4 Transformation events created at public research institutions related to
quality-of-life categories
Trait No. of eventsa
Insect resistance 35
Lepidoptera 35
Disease resistance 82
Bacteria 8
Fungi 21
Viruses 53
Abiotic stress tolerance 11
Drought 7
Salinity 4
Quality improvement 15
Nutritional and other 9
Enhancing shelf-life 6
Other 9
Vaccines 9
Total in this table 152
Total number of reported events 201
Percentage of all events related to
quality-of-life traits 75.6%
Figure 6 Projected dissemination plans for final
research transformation events (in GM crops).

NATURE BIOTECHNOLOGY VOLUME 23 NUMBER 1 JANUARY 2005 33
FEATURE
resources will determine if local resources or
adapted genes need IP protection. Benefit dis-
tribution, accounting for the success in trans-
forming local genetic resources, can form the
basis for agreements between public institutes,
farmer organizations and commercial produc-
ers18. Agreements can establish ownership
among providers of transgenes (and the cost
of their research) by equalizing investments
with time and innovation provided by develop-
ing countries creating combinations of genes
in localized crops or genetic resources19. Such
decisions on ownership are made carefully to
ensure an equitable arrangement between poor
country institutions20 and those supplying new
technologies. Our data offer many examples
where further investigation into ownership
would be of benefit, as abilities grow for incor-
porating privately developed genes into crops
of local value.
Local and multinational companies could
play a key role for specific local GM crops,
given their experience in commercial devel-
opment and regulatory information, includ-
ing environmental and food safety studies.
However, examples of successful public-private
partnerships in plant biotech are still rare, even
at international research centers21.
Creating efficiencies and competencies.
Although limited collaboration does occur
between developing countries and Western
companies (Tabl e 3 ), the study reveals that
developing countries did not forge a single
(‘South-to-South’) collaboration among
themselves. Contacts with other countries of
economic parity would create efficiencies by
sharing knowledge on specific crops, traits and
regulatory dossiers. For example, by using data
on genes and phenotypes under study (Ta b l e
2), countries could meet and assemble data
and experience on specific genes and their
constructs, making collected and relevant
information available to their respective regu-
lators. Scientists and regulators from develop-
ing countries can also meet to discuss specific
crops, where common transformation events
are occurring.
Working from either specific crops or traits,
joint studies can also highlight partnership
models (or the lack of them) and address needs
best suited for such collaboration. The same
type of consultation can occur by examining
crops at a particular stage in their regulation
(Fig. 4), their required safety information and
results from efficacy and safety trials. Such
knowledge is valuable when selecting trans-
genes, considering regulatory requirements
and determining which genetic resources are
available or needed.
The bottom line. Although some commer-
cially developed GM products have a role to
play, GM crops developed by public research
institutes should be most relevant to local
needs in poor countries. Paradoxically, because
they are novel, locally developed products pose
unique challenges for institutes seeking regula-
tory approval, and gaining approval can be one
of the biggest obstacles facing public GM crops
in developing nations. In contrast, commercial
GM crops preapproved in Western markets are
more successful in gaining approvals in devel-
oping countries.
Demand for GM products by local farmers
combined with the established regulatory and
production track record of Western products
sets the stage for interest in using GM crops in
developing nations. This implies farmers may
take advantage of options to grow locally unap-
proved Western products, thus avoiding licens-
ing costs and IP issues. At the same time, locally
produced GM crops remain in development
and do not reach the same farmers, meaning
their impact is yet to be seen.
ACKNOWLEDGMENTS
The author acknowledges contributions to this
paper from P. Zambrano of IFPRI, and all those
participating in the initial Next Harvest study.
Special thanks to the National Agricultural
Biotechnology Council Meeting 16, where this
data and ideas were initially presented, and for the
support of Alan Wildeman. Financial and technical
support from The Netherlands DGIS in The Hague,
The Netherlands, the Swiss Agency for Development
Cooperation (Bern) and the United Kingdom’s
Department for Internal Development (London).
1. De Greef, W. The Cartagena Protocol and the future of
agbiotech. Nat. Biotechnol. 22, 811–812 (2004).
2. Atanassov, A. et al. To Reach The Poor. Results from
the ISNAR-IFPRI Next Harvest Study on Genetically
Modified Crops, Public Research, and Policy Implica-
tions. EPTD Discussion Paper 116. (International Food
Policy Research Institute, Washington, DC, 2004).
3. Asian Development Bank. Agricultural Biotechnol-
ogy, Poverty Reduction, and Food Security. A Working
Paper. (Asian Development Bank, Manila, the Philip-
pines, 2001). http://www.adb.org/Documents/Books/
Agri_Biotech/agribiotech.pdf.
4. Cohen, J.I. Harnessing biotechnology for the poor:
challenges ahead for capacity, safety, and public
investment. J. Human Dev. 2, 239–264 (2001).
5. United Nations Economic Commission for Africa. Har-
nessing Technologies for Sustainable Development.
(United Nations Economic Commission for Africa,
Addis Ababa, 2002).
6. Josling, T., Roberts, D. & Orden, D. Food Regulation
and Trade. Towards A Safe and Open Global System.
(Institute for International Economics, Washington,
DC, USA, 2004).
7. Kameri-Mbote, P., Wafula, D. & Clark, N. Public/Private
Partnerships for Biotechnology in Africa. ACTS/Bio-
Earn Occasional Paper. (African Center for Technology
Studies, Nairobi, Kenya, 2001).
8. Hall, A. et al. Public-private sector interaction in the
Indian agricultural research system: an innovation sys-
tems perspective on institutional reform. in Agricul-
tural Research in an Era of Privatization (eds. Byerlee,
D. & Echeverria, R.G.) 155–176 (CABI International,
Wallingford, UK, 2002).
9. Food and Agricultural Organization. State of Food and
Agriculture (SOFA). Report for 2004. (FAO, Rome,
2004).
10. Watanabe, K.N., Taeb, M. & Okusu, H. Putting Carta-
gena into practice. Nat. Biotechnol 22, 1207–1208
(2004).
11. Mugoya, C. & Bananuka, J.A. (eds). Resource Book for
Implementation of Biosafety in East Africa (Bio-EARN,
Kampala, Uganda, 2004).
12. Wong-Anan, N. Thais lift ban on GMO planting, will
regulate trials. USA Today August 23, 2004.
13. Puttajanyawong, T. Thai cabinet overturns GMO
approval. Reuters. 2004-08-31 (2004).
14. Phillips, P.W.B. Policy, national regulation and inter-
national standards for GM foods. in Biotechnology and
Genetic Resources (eds. Pardey, P. & Koo, B.). Brief 1,
1–5 (IFPRI, Washington, DC, USA, 2003).
15. Cohen, J.I. & Paarlberg, R. Unlocking crop biotechnol-
ogy in developing countries—a report from the field.
World Development 32, 1563–1577 (2004).
16. Swaminathan, M.S. Report of the Task Force on the
Application of Biotechnology In Agriculture (Govern-
ment of India, Ministry of Agriculture, Krishi Bhawan.
New Delhi, India, 2004)
17. Abdallah, R. & Bamwenda, G.R. (eds.) Initiating Agri-
cultural Biotechnology in Tanzania (Tropical Pesticides
Research Institute, Arusha, Tanzania, 2004). http://
www.fao.org/biotech/inventory_admin/dep/default.
asp
18. Mahoney, M.M, V. Henson-Apollonio, and H. Hambly
Odame. 2004. Strategies for management of intel-
lectual property in developing countries and the role
of farmers’ associations. ISNAR Briefing Paper 78.
ISNAR Program, Addis Ababa, Ethiopia.
19. Goodman, M.M. and M.L. Carson. Myth vs. reality:
corn breeding, exotics, and genetic engineering.
Annual Corn Sorghum Research Conference Proc. 55,
149–172 (2000).
20. Pardey, B. et al. Intellectual property and develop-
ing countries: freedom to operate in agricultural bio-
technology. in Biotechnology and Genetic Resources
(eds. Pardey, P. & Koo, B.) 1–6 (Washington, DC, USA,
2003).
21. Spielman, D. & Von Grebmer, K. Public-Private Part-
nerships in Agricultural Research: an Analysis of Chal-
lenges Facing Industry and the Consultative Group on
International Agricultural Research EPTD Discussion
Paper No.11. (International Food Policy Research
Institute, Washington, DC, USA, 2004).
22. FAO-BioDeC. Database on Biotechnology in Developing
Countries (2003). http://www.fao.org/biotech/inven-
tory_admin/dep/default.asp
23. FAO STAT Item Codes. http://apps.fao.org/faostat/form
?version=ext&collection=Production.Crops.Primary&D
omain=Production&language=EN&servlet=1&axis=ite
m&xsl=areareflist