Page 1
Charting the Road Map for Long-Term USDA Efforts in Agricultural Animal Genomics:
Summary of the USDA Animal Genomics Workshop – September 2004
R. D. Green
, M. A. Qureshi
, and J. A. Long
USDA-ARS, Beltsville, Maryland, USA
USDA-CSREES, Washington, District of Columbia, USA
Agricultural animal research has been immensely successful over the past century in
developing technology and methodologies that have dramatically enhanced production
efficiency of the beef, dairy, swine, poultry, sheep, and aquaculture industries. Most of this
research effort has been conducted within the broad disciplines of genetics, physiology, and
nutrition, but has become increasingly narrowly defined into multiple sub-disciplines over
time. While a vast body of knowledge has been generated from this effort, it has now
become clear that the majority of the potential for future improvements in animal production
efficiency, quality of animal products, and animal health lies in the elucidation and
understanding of interactions of the various components of the biology of the animal in
concert with all of the parameters of the production environment. To begin to fully
understand these interactions, a redirection of the traditional “reductionist science” approach
to a “systems biology” approach is required.
In the past two decades, molecular biology has changed markedly the face of agricultural
animal research, primarily in the arena of genomics and the relatively new offshoot areas of
functional genomics, proteomics, transcriptomics, metabolomics and metagenomics.
Publication of genetic and physical genome maps in the past 15 years has given rise to the
possibility of being able finally to understand the molecular nature of the genetic component
of phenotypic variation. While quantitative geneticists have been remarkably successful in
improving production traits, genomic technology holds potential for being able to lead to
more accurate and rapid animal improvement, especially for phenotypic traits that are
difficult to measure.
Recently, the agricultural research community has been able to capitalize on the
infrastructure built by the human genome project (Collins et al., 2003; Gibbs et al., 2005) by
sequencing two of the major livestock genomes (Gallus domesticus (Wilson et al., 2004;
Andersson et al., 2004) and Bos taurus (Gibbs et al., 2002)). The 2005 calendar year is truly
unprecedented in the history of agricultural animal research since annotated draft genome
sequences were completed for chickens and cattle and draft genome sequences will be in
progress for swine in early 2006. We now have in place the foundation of a powerful
toolbox for understanding the genetic variation underlying economically important and
complex phenotypes.

Page 2
Over the past few years, new challenges have emerged for animal agriculture. Enhancements
in production efficiency have not come without some negative side effects on animal well-
being and longevity in production environments, including losses in reproductive efficiency,
increased stress susceptibility, increased animal waste issues, and increased susceptibility to
animal metabolic and infectious diseases. When considered in concert with increased
societal concerns in the areas of natural resource conservation and protection, animal welfare,
and food safety, it is clear that publicly supported agricultural research must be focused on
enhancing the functionality and well-being of livestock and poultry in environmentally
neutral production systems in the future.
Realizing the great potential for animal genomics to address these and other issues, a
workshop was convened by the U. S. Department of Agriculture (USDA) in Washington, DC
in September of 2004. The workshop was entitled “Charting the Road Map for Long Term
USDA Efforts in Agricultural Animal Genomics”. The objective of this paper is to
summarize the proceedings of the workshop and the resulting recommendations.
The 20
century was an immensely prolific time in developing methods to enable genetic
improvement in livestock and poultry. In the first half of the century, geneticists who were
busy trying to understand the nature and behavior of chromosomes converged with
biometricians who had developed considerable statistical theory that could be used to
describe variation observed in genetically defined populations of animals and plants. These
two fields coalesced into what became known as “quantitative genetics” around the time of
World War II. Over the following forty years, sophisticated genetic prediction
methodologies were developed for the dairy, beef, swine, sheep, and poultry industries for a
number of the economically important production traits measured on seedstock populations
of animals. This body of work was based upon the assumption that differences observed in
performance phenotypes could be attributed to underlying heritable genotypic differences. It
was assumed as well in these approaches that for most of the economically important traits
(e.g., milk yield, growth rate, meat yield, etc.) a large number of unknown genes contributed
to this heritable variation, which led to breeding value predictions based on an infinitesimal
gene model “black box” approach. An entire “genetics industry” developed around this
framework and created many tools, such as predicted differences in dairy cattle and expected
progeny differences in beef cattle. Even though it was probabilistic in nature, this approach
has been highly successful in allowing directed genetic change to occur in all of these
species. An excellent example of this approach is the coupling of genetic selection and
efficacious use of artificial insemination over the past 50 years that led to a more than 100%
increase in annual milk yield per cow.
In the mid-1980s, a new window of opportunity opened in livestock production science. In
1986, the new term “genomics” was coined to refer to the new technologies that were
developed and applied to the study of mammalian DNA, such as the application of bacterial
restriction endonucleases for rudimentary visualization of differences in the sequence of
DNA in particular chromosomal locations through “restriction mapping”. This was followed

Page 3
quickly by the development of the polymerase chain reaction in 1987 that opened up an
entirely new world for the study of differences in the DNA sequence of animals. Coupled
with the discovery of short tandem repeat DNA markers, PCR became a powerful tool that
quickly allowed the development of genetic maps of the livestock genomes, primarily based
on linkage of microsatellite DNA markers.
In 1990, an Allerton Conference entitled “Mapping Domestic Animal Genomes: Needs and
Opportunities” was hosted by the University of Illinois. This conference provided the first
opportunity for scientists, producers, industry, and government representatives to come
together to discuss how emerging molecular technologies could be employed to bring about
major innovations for animal agriculture. Participants at this workshop recommended to
USDA that genetic maps be developed to a 20 cM saturation level for each of the
agriculturally important species (cattle, swine, poultry, fish, and horses). This
recommendation was implemented and resulted in the publication of a number of important
genetic linkage maps in the mid-1990s.
With the initial genetic maps in place, a second Allerton Conference, entitled “Genetic
Analysis of Economically Important Traits in Livestock”, was convened in 1996 to address
capitalizing on animal genomics research. The workshop focused on statistical approaches to
mapping complex quantitative traits and discerning how any DNA markers identified
through such mapping could be used in selection programs. By this time, a number of
research groups around the world had developed resource family populations that were being
employed, using the previously developed linkage maps, to identify regions of the genome
appearing to harbor genes giving rise to phenotypic variation in complex traits (so-called
Quantitative Trait Loci (QTL)). Once DNA markers anchoring these QTL regions were
identified, it was postulated that “marker-assisted selection” could be used to make directed
genetic change in the desired traits using this technology. The primary recommendation of
the Allerton II workshop was a call for building the research infrastructure necessary to
enable researchers to identify important genes that control economically important traits and,
eventually, gain an understanding of the function of individual genes and their interactions
across the genome.
By the end of the century, everyone recognized that more genomic tools and resources were
necessary for the fulfillment of the promise of livestock and poultry genomics. Even though
a large number of putative QTL had been identified for a wide spectrum of traits, only a
handful of simply inherited traits had been elucidated through this approach. In all of these
successful cases, the fine mapping of the identified genes had relied on comparative mapping
approaches to make use of the denser information available in the human, mouse, and rat
maps. Despite having some improved tools, such as bacterial and yeast artificial
chromosome libraries, it became clear that without the availability of the whole genome
sequence as a scaffold from which to work, the time and expense of fine QTL mapping was
much greater than initially envisioned. Fortunately, new high-throughput technologies were
being developed that made the sequencing of whole genomes more practical, efficient, and
cost effective. The human genomics research community quickly recognized this
opportunity and the government and privately funded human genome sequencing projects
were launched.

Page 4
As the 21
century began, and the human genome moved toward an initial draft sequence,
other new technologies also became available that allowed livestock and poultry researchers
for the first time to move into large-scale gene expression studies. By coupling expressed
sequence tags with new microarray technologies, researchers were able to visualize changes
in levels of expression of hundred or thousands of genes in specific tissues under a wide
variety of conditions. This began to broaden genomics research into the functional realm and
initiated open discussions on how genomics might be used to bridge various disciplines into a
“systems biology” framework.
In 2001, the Alliance for Animal Genome Research was formed by a group of universities,
private industry parties, producer groups, and scientific societies to advocate for public
funding for domestic animal genomics. This group was successful in working with the
National Academy of Sciences to organize a public workshop held in 2002 entitled
Exploring Horizons for Domestic Animal Genomics” with the goal of identifying research
goals and public and private funding needs (National Academy of Sciences, 2002). There
was overwhelming consensus at the workshop that funding should be identified to produce
high-coverage, draft genome sequences of the major domestic animal species (cattle,
chicken, swine, dog, and cat) for deposit into the public domain databases. NHGRI had
previously established a process for prioritizing species for sequencing based upon the ability
of a species to better inform annotation of the human genome sequence through evolutionary
comparisons. The workshop participants felt that these species would be excellent
candidates to meet that objective in addition to the fact that they had been used heavily as
biomedical models and were important agricultural or companion animal species.
Furthermore, it was recommended that there would need to be appropriate scaling-up of
bioinformatics resources to make effective use of the volumes of information that would
result from the genome sequencing projects. Based upon the experiences of the National
Plant Genome Initiative, it was also recommended that funding for such large-scale projects
would need to come from a variety of sources, including the U.S. Federal government,
private industry, and international partners.
In July of 2002, a third Allerton Conference entitled “Beyond Livestock Genomics” brought
together leading investigators from a broad spectrum of disciplines (genetics, physiology,
reproduction, animal health, and nutrition) to develop an initial plan for the full utilization of
genomic information to promote animal health and productivity, while more broadly
contributing to the greater life sciences. The overarching recommendation from this
workshop was that additional basic research was needed to identify genomic mechanisms and
novel genes / proteins in a variety of tissues under a variety of environmental conditions
(Hamernik et al., 2003). Functional genomics was recognized as the vehicle for capitalizing
on the investment of obtaining whole genome sequence information. The need to increase
bioinformatics infrastructure and teaching and outreach efforts in animal genomics was
recognized also.
In response to a request by USDA Undersecretary Joe Jen, a new Interagency Working
Group (IWG) on Domestic Animal Genomics was chartered in September of 2002 by the
U.S. National Science and Technology Council with the mission of enhancing

Page 5
communication and awareness of the importance of livestock and companion animal species
of importance to the food and agriculture system; increasing leverage of Federal investments
in large-scale genome sequencing and genome analysis across government agencies;
positioning the food and agriculture system as a critical element of the national genomics
program; enhancing dialogue and cooperation among Federal agencies, universities, and
industry in the nation; and promoting international cooperation on domestic animal genomics
research. The membership of the IWG has included representatives from the Department of
Agriculture (USDA), Department of Energy (DOE), Food and Drug Administration (FDA),
National Institutes of Health (NIH), National Science Foundation (NSF), Office of Science
and Technology Policy (OSTP), Office of Management and Budget (OMB), Department of
Homeland Security (DHS) and the U.S. Agency for International Development (USAID).
The IWG subsequently identified the following broad strategic goals:
• Bring into place the programmatic elements needed to advance the study and
understanding of domesticated animal genomes, including large-scale DNA
sequencing; functional characterization of expressed genes (functional genomics);
tools for data storage, analysis and visualization (bioinformatics); and study of
similarities among genomes of different species (comparative genomics).
• Leverage the national infrastructure for large-scale DNA sequencing that has been
established for the Human Genome Project and other vertebrate and model organism
• Advance and utilize the enabling tools and infrastructure of functional genomics and
bioinformatics to enhance the understanding not only of basic science and disease
mechanisms, but also to address critical agricultural missions, including animal health
and well-being, food safety, and human nutrition.
• Ensure that genomics data are freely available in the public domain and genomics
reagents and resources are available to the public.
• Increase the training opportunities for genomics and bioinformatics at all levels of
• Coordinate and encourage international cooperation to achieve these goals.
The IWG determined that large-scale sequencing, data management and bioinformatics, and
functional genomics were the specific goals to be achieved in fiscal years 2003 to 2007. The
IWG called for, among other things:
• Large-scale sequencing to produce draft genome sequences (8-fold sequence coverage)
of honeybee, chicken, dog, cattle, swine, and cat species.
• Data management and bioinformatics to specifically support agriculturally important
species, including significant improvements in data management and analysis
software, allow for greater data accessibility and secure long term maintenance,
increase capabilities to deal with rapidly accumulating data complexity as databases
include functional information, and provide more powerful tools to mine large
• Recognition that an increase in data for livestock genomes requires a concomitant
investment in functional genomics to support genome annotation, the study of gene
regulation and expression, and species evolutionary relationships.

Page 6
Since 2002, considerable progress has been achieved towards the goal of placing whole
genome sequence and associated tools into the public domain for high priority domestic
animal species. Annotated draft sequences have been published for the honeybee, chicken,
and dog genomes and the bovine genome sequencing project has reached >6-fold coverage
and is entering the final gene prediction and annotation phase. Lighter coverage sequencing
of the cat genome has been completed and BAC-skim sequencing of the swine genome was
launched in January 2006. Developed concomitantly with these genome projects has been a
suite of associated tools including EST libraries, BAC maps, integrated physical and linkage
maps, full-length cDNA libraries, microarrays or gene chips, and identification and
validation of a large number of single-nucleotide polymorphism markers. All of these efforts
have required leveraging of efforts between agriculture and the biomedical sciences, as well
as unprecedented partnerships between U.S. Federal research agencies, international groups,
universities, and private industry.
In early 2004 as the sequencing goals of the IWG appeared to be within reach, further study
of how to best address the remaining two areas of greatest importance – bioinformatics and
functional genomics – was warranted. Specifically, the charge was given by the IWG to the
USDA to evaluate how programs in these areas should be developed further to allow full
utilization of annotated genome sequences and associated tools.
Life sciences research activities in the USDA are administered by two separate agencies.
The Cooperative State Research, Education, and Extension Service (CSREES) funds
extramural research efforts conducted primarily at land grant and 1890s universities. The
Agricultural Research Service (ARS) is the intramural research arm of the USDA and funds
long-term, high-risk research on an ongoing basis in its 108 labs throughout the U.S. In
fiscal year 2004, USDA funding for animal genomics research totaled $46.4M (ARS -
$22.7M and CSREES - $23.7M).
The USDA Animal Genomics Workshop, as called for by the IWG, was designed to
facilitate open input and discussion from leading scientists working in the field of animal
genomics for USDA administrators. Participating scientists were selected to reflect a balance
of funding sources (16 from CSREES and 18 from ARS), species of primary interest
(balance between poultry, swine, cattle, sheep, and aquaculture), and area of research
emphasis or expertise (gene mapping, bioinformatics and statistical genetics, functional
genomics). In addition to a number of program administrators from ARS and CSREES,
colleagues from NIH, NSF, and DHS also participated in the workshop.
The workshop was organized in three modules. Each module consisted of presentations by
an invited panel of speakers followed by three simultaneous breakout groups to discuss the
long-term needs in that area. Group reports were then assimilated into a consensus set of
recommendations emanating from the event.

Page 7
Structural Genomics Priorities in Domestic Animal Genomics. The opening session
focused on the structural genomics needs facing animal genomics researchers today. Noelle
Cockett, Utah State University, provided an overview to set the stage for further discussion.
Generally, scientists have approached genomics primarily by building structural genomics
resources, with ventures into functional genomics observed only in more recent years. The
animal genomics research community has been successful in prioritizing needs in annual or
semi-annual meeting venues, such as the International Plant and Animal Genome Workshops
and the International Society of Animal Genetics. Through such international collaborations
and efforts, linkage and comparative maps for all livestock species were made available. The
recent and ongoing development of whole genome sequence maps of the chicken, honeybee,
dog, and cattle species is a major step forward. Other important agricultural species
including swine, several aquaculture species, and sheep are attempting to develop funding
resources to enter the sequencing pipeline. Single-nucleotide polymorphism (SNP) based-
maps now being developed from the chicken and cattle whole genome sequencing projects
will be of enormous value in evaluating genetic diversity, fine mapping of QTL, and
development of DNA-based animal identification systems. While the current trend toward
internalization of genomics research in private companies indicates the potential value of
genomic tools, it also was pointed out as a major concern. There was a consensus that we
must complete the basic genome infrastructure for all major species and deposit such
information in public databases, as this was viewed as absolutely essential to facilitate rapid
discovery and the development of commercially usable technologies for agricultural and
biomedical sciences and industries
A major advantage of using agricultural animal species in genomics research is the
widespread availability of large, pedigreed research animal populations. Many of these
populations have been in existence for fifty years or more and have been phenotyped widely
for a variety of economically and biologically important characteristics. In the past two
decades, a number of sub-populations were set up as resource families for use in QTL
detection and subsequently for validation of putative QTL. Participants agreed that it is
imperative in the post-genome sequencing era that the value of these populations, and tissue
repositories derived from them, be recognized and supported.
Participants agreed that animal genomics is poised to impact several avenues of animal
production, life sciences, and biomedical research, but physical and financial resources are
crucial to capitalizing on past investments. The utilization of resources and human capital
must, however, be carefully directed toward achieving outcomes and deliverables that are
measurable in application, promote rapid commercialization, and enhance education of the
public and the next generation of scientists. The need for a cohesive, comprehensive long-
term plan for all of USDA’s research efforts in genomics was evident at the workshop.
Further integration of the efforts of CSREES and ARS appears warranted to achieve the
greatest return on investment.
Specific Recommendations from the Structural Genomics Module:
1) Sequence the swine genome to a minimum of 6-fold coverage for deposit into the
public domain.

Page 8
2) Obtain BAC maps and 2-fold sequence coverage for the catfish, goat, horse, salmon,
trout, and turkey genomes.
3) Develop comprehensive full-length cDNA libraries to allow functional annotation to
be achieved to acceptable levels for each of the genome assemblies listed under #2.
4) Complete integration of genetic linkage, radiation hybrid, and physical maps should
be achieved for each genome listed under #2.
5) Discover and validate SNP markers and develop haplotype maps for all species to
increase the density of maps for fine mapping of QTL and eventual “whole genome
6) Develop standardized population and phenotype resources for each of the species.
a. Preserve long-term, unique, experimental animal populations to capitalize on
their value in functional genomics research and further develop and maintain
diverse animal resource families.
b. Couple these animal populations with genotypic and phenotypic information
and obtain funding support for appropriate long-term tissue repositories for
tissue cultures, DNA and RNA.
c. Explore options to bring the agricultural animal genomics community in line
with the laboratory mouse and rat communities (i.e. the Jackson Labs model).
[The National Animal Germ Plasm Program, currently administered by
USDA, may provide a foundation upon which to build for this function. The
IWG should study this carefully to avoid any unnecessary duplication of effort
and resources across Federal agencies.]
Long-Term Challenges in Making Use of Genome Sequence Information through
Functional Genomics. The second module of the workshop was an open discussion of the
challenges facing agricultural animal genomics researchers in capitalizing on the structural
genomics infrastructure through downstream applications in functional genomics,
proteomics, metabolomics, and metagenomics. Three working groups were assigned in
advance of the workshop to develop presentations representing the genomics research
communities working with poultry, swine, and ruminants.
Jerry Dodgson, Michigan State University, presented an overview of chicken genome
research and associated challenges. With the availability of the chicken physical maps,
ESTs, microarrays, and, most importantly, the release of the draft genome sequence in 2004,
the chicken became the first avian genome sequence available to scientists worldwide. It was
stated that the domestic chicken has retained enormous genetic diversity, based on
comparative SNP-based studies of three chicken breeds (broiler, layer and Chinese silkie)
relative to the Red Jungle Fowl used for the genome sequence. Chickens possess an
abundance of quantitative variation in production and disease resistance traits and are a
unique biomedical model in addition to being a leading source of high quality animal protein
worldwide. For these reasons, a case was made for finalizing the draft genome sequence of
chicken. The chicken genome community will face grand challenges similar to those faced
by the human genome community in the post-genome sequencing era, including: 1)
identifying the structural and functional components encoded in the genome; 2) elucidating
the genetic networks and protein pathways and their relation to phenotypes; and 3)
understanding and applying the heritable variation in the genome.

Page 9
Harris Lewin, University of Illinois at Urbana-Champaign, presented the long-term
challenges associated with making use of genome sequence information through functional
genomics in ruminants. An improved understanding of the genomic basis for traits of
economic importance to the dairy, beef and sheep industries was identified as an important
goal. Research problem areas where functional genomics would contribute include, but are
not limited to, embryonic development (pre- and post- implantation), lactation (efficiency,
composition and product quality), wool growth, muscle growth and meat quality, feed
efficiency, immunobiology of infectious diseases, and animal well-being. Ultimately, the
selection of candidate genes and the identification of allelic variation associated with the
phenotypes are important products of functional genomics. For cattle, genomic resources
available to accomplish the functional genomics goals include linkage and radiation hybrid
maps with a large number of markers, BAC libraries, and physical maps developed through
the International Bovine BAC Map Consortium, and whole genome sequence information.
Germplasm repositories, animal resources, QTL maps, ESTs, microarrays and associated
databases are in place and available as additional resources to accomplish the functional
genomics goals. The “grand challenges” for the ruminant genomics community are: 1)
functional annotation of cattle (and other ruminant) genes; 2) complete description and
understanding of cellular pathways (e.g., metabolism, proliferation, differentiation, cell-cell
interaction); 3) genomic-environment interaction (e.g., developmental pathways, abiotic
stresses such as heat, cold, and drought, nutritional genomics, and infectious diseases); and 4)
the development of an encyclopedia of economic trait loci. A need for additional biological
resources (e.g., tissue banks, animal germplasm, cell lines), genomic technologies (e.g.,
RNAi, genotyping services, cloning and transgenics) and integrative databases and
informatics was identified.
Larry Schook, University of Illinois at Urbana-Champaign, presented on behalf of the swine
genomics research community. It was emphasized that to answer key biological questions, it
is essential to have a whole genome sequence to harness comparative functional genomics
across species. A minimum 6-fold coverage of the swine genome was recommended for
carrying out a high quality, functional genomics program and to remain compatible with the
NIH genomics programs. He outlined the International Swine Genome Sequencing
Consortium’s efforts in identifying the needs and resources for the swine genome sequencing
initiative. A timetable was presented and it was emphasized that the international researchers
and industry leaders are in agreement that a swine genome sequence is needed and the effort
is timely.
General Discussion. Downstream or post-genomic applications, such as functional
genomics, proteomics and metabolomics, clearly are the areas where agricultural species will
benefit from the genome sequencing research investments. These benefits have begun to be
realized with the completion of the human genome sequence; for example, over 40 genes
have been identified subsequently for a variety of conditions, including macular degeneration
(Yamagishi et al., 2005), cleft palate (Frebourg et al., 2005), lymphoproliferative disease
(Nichols et al., 2005), mental retardation (Jensen et al., 2005), and testicular cancer
(Diederichs et al., 2005). For both human and agricultural species, the post-sequencing
challenge will be to understand the operation and function of genomic information. In

Page 10
particular, the primary issue for agricultural species will be translating the respective genome
sequences into enhanced productivity of the phenotypes they control or influence (e.g.,
disease resistance, behavior, growth, product quality, reproduction).
The post-sequencing era will move rapidly from crudely defined genomic relationships with
phenotypes, such as QTL, to a rapid dissection of those relationships in the context of true
functional genomics. Some examples of QTL that should progress rapidly from chromosomal
localization to industrial application include meat quality and product yield in beef cattle,
milk production and mastitis resistance in dairy cattle, litter size and uterine capacity in
swine, product yield and parasite resistance in sheep, and coccidia resistance in poultry. The
availability of genome sequences for agricultural species will enhance significantly fine
mapping of individual genes in two key ways. First, an exponential increase in the numbers
of SNPs distributed throughout the linkage maps will enable fine mapping of QTL at a level
previously not possible. For example, poultry genomics is poised to realize this benefit with
the placement of some 3 million SNPs on a 1.2 Gb genome. Second, comparative genomics
will increase the likelihood of QTL identification by virtue of the highly conserved regions of
genes throughout mammalian species (e.g., myostatin gene responsible for double-muscling
condition in cattle [Grobet et al., 1997; Casas et al., 1998; Yang et al., 2001]).
The majority of economically important traits exhibit complex or multifactorial inheritance
patterns that are influenced by environmental factors; therefore, the principal challenge is not
simply detecting the QTL, but rather unraveling the genes and the regulatory elements that
control gene expression (Andersson and Georges, 2004). This will require the integration of
numerous resources, including genetic and physical maps, QTL markers, EST libraries,
microarrays and the whole genome sequence to delineate the molecular mechanisms that
control complex biological systems. Agricultural species have an advantage in that
phenotypes are well characterized and diverse because they have been closely monitored and
specifically modified through selected breeding.
Expression profiling of large numbers of genes across diverse tissues, populations, and
environmental states also will use increasingly sophisticated quantification platforms. For
example, the expression of literally thousands of genes can be studied simultaneously already
using DNA chips or microarrays. The molecular biologist will be able to bypass traditional
laborious processes, such as screening BAC libraries, and instead clone genes “in silico”
(Wong, 2004). Proteomic technologies, including new developments in mass spectrometry
and database searching, are leading to rapid advances in monitoring genome activity at the
protein level. We can expect further advances in understanding the structural biology of
proteins when comparative and evolutionary approaches to sequencing are utilized.
Proteome analysis will elucidate groupings of genes that regulate metabolic pathways.
Additionally, by following gene expression fluctuations over time and in response to specific
signals, the position occupied by the protein end product of a particular gene, relative to
others in metabolic and signaling pathways, can be inferred (Roberts, 2001). It follows, then,
that fields, such as metabolomics, will allow genomic characterization of “systems” of
proteins and their applications to animal health and nutrition, as well as human nutrition and
obesity. Whereas genes and proteins set the stage for what happens in the cell, much of the

Page 11
actual activity is at the metabolite level: cell signaling, energy transfer, and cell-to-cell
communication are all regulated by metabolites (Schmidt, 2004).
New technologies will continue to be developed at a rapid pace to improve both the precision
and efficiency of the various ‘omics’ approaches. For instance, the phenomenon of RNA
interference (RNAi) has evolved rapidly into a powerful technique to silence gene expression
in eukaryotic cells. Agricultural researchers have begun to use this technology to study gene
function in porcine granulosa cells (Hirano et al, 2004) and bovine ooctyes (Paradis et al.,
2005). Because RNAi technology can be used to knock out genes across a genome, having
the complete genome sequence will greatly improve identification of ‘targets’ (proteins) for
existing drugs. For example, parasitologists at CSIRO Livestock Industries are using this
approach in an effort to control internal and external parasites of cattle and sheep. Another
emerging technology, metagenomics, is poised to develop rapidly and have profound impacts
on functional genomics research in agricultural species. Metagenomics is a new field
combining molecular biology and genetics in an attempt to identify and characterize the
genetic material from environmental samples and apply that knowledge. Genetic diversity is
assessed by isolation of DNA followed by direct cloning of functional genes from the
environmental sample. The metagenomics field was pioneered when researchers used whole
genome shotgun sequencing to sequence microbial populations en masse from the Sargasso
Sea (Venter et al., 2004). It is not hard to envision application of this technology to
ascertain the microbial populations of the bovine rumen or porcine intestine, for example,
and how the dynamics/interactions among bacterial and protozoan species create a unique
microenvironment that promotes growth.
Although the field of transgenic animal production is not new in comparison to RNAi and
metagenomic tools, this is an example of existing technology where needed improvements
will accelerate and culminate in the development of model animal systems using livestock
species. Larger domestic animals are valid biomedical research models by virtue of their
anatomical and physiological similarity to humans. For example, the retina of the rhodopsin
transgenic pig (Petters et al., 1997) shares many cytological features with human retinas
exhibiting retinitis pigmentosa, a degenerative loss of cone photoreceptors that gradually
leads to blindness. Most recently, this transgenic animal model has been used to develop
surgical transplantation of normal neuroretinal grafts (Ghosh et al., 2004). The widespread
use of existing transgenic domestic animal models has been limited by the relatively low
success rates of nuclear transfer and cloning. To illustrate the low efficiency of producing
transgenic animals, consider that a minimum of 1,200 microinjected eggs were required to
produce one transgenic sheep, goat, or cow and that only about 50% of offspring express the
transgene (Wall et al., 1997). A prime example of the enormous potential of transgenic
technology is the recent production of transgenic dairy cattle with resistance to mastitis (Wall
et al., 2005). The production of 8 transgenic calves, however, required embryonic transfer of
927 good quality blastocysts that were created from over 4,000 nuclear donor cells. Using
similar technology, dairy cattle also present a great potential for producing amounts of
therapeutic proteins secreted into milk. Likewise, eggs from transgenic hens are a potential
high-throughput mechanism for production of therapeutic proteins. Additionally, the avian
transgenic system may confer post-translational glycosylation processing more similar to
humans than other species currently used for transgenic production of proteins. It is clear

Page 12
that gene transfer technologies will have renewed focus in the post-sequencing era of
Perhaps the most intriguing example of new technology development on the heels of genome
sequencing was the call for proposals from the National Human Genome Research Institute
(NHGRI) in 2004 seeking the next generation of technologies that would reduce the cost and
increase the throughput of DNA sequencing. In short, the goal of NHGRI is to lower the cost
of sequencing one individual’s genome (human or animal) to $1000 (USD). Once in place,
these technologies will further revolutionize the post-sequencing era for agricultural species.
With all of the expected and rapid increases in knowledge in the near future, it is imperative
that the methodology for defining phenotypes be clear and standardized. The systematic
classification and characterization of phenotypes is essential for ultimately mapping the
genes responsible for normal and abnormal development and physiology. More importantly,
any search for mutations or altered functional expression depends on phenotypic screening
and the ability to detect variation from normal. The challenge, then, is to develop efficient,
systematic, and comprehensive phenotypic screening procedures and tools that will permit
comparison among laboratories. For example, the current phenotypes of highly pathogenic
avian influenza (HPAI) were formulated over 10 years ago when the only virus known to
have mutated to virulence was the HPAI responsible for the 1983–84 Pennsylvania epizootic
(Alexander, 2002). Cumulative evidence, however, suggests that HPAI viruses actually
arose from low-pathogenicity avian influenza (LPAI) H5 or H7 viruses infecting poultry after
spreading from free-living birds. At present, it can only be assumed that all H5 and H7
viruses have this potential and mutation to virulence is a random event. Therefore, the longer
the presence and greater the spread in poultry, the more likely it is that HPAI virus will
emerge (Alexander, 2002). This example illustrates how major research efforts in
phenotypic screening are needed to characterize traits that have been difficult to measure
until now.
Concomitant with the advent of functional genomics, the types and amounts of data that need
to be stored in databases have changed dramatically. Many types of information that were
previously collected on an ad hoc basis now need to be stored in a more structured manner.
Additional data sets for gene expression, proteomics, and protein-protein interactions are
growing increasingly complex. To analyze the data computationally in an efficient manner,
there is a need for consistency between expressions in different phenotypic domains as well
as in different species. The term “phenotype” can be used in different ways in different
fields in biology and by different researchers in those fields. It may mean anything from the
complete set of phenotypic attributes (traits) that describe an individual to a single
phenotypic attribute that distinguishes an individual from other, “normal” individuals
(Gkoutos et al., 2004). The development of phenotypic ontologies for livestock is critical to
our ability to connect heterogeneous data types back to animal. It would be best to define the
ontology in a proactive manner so that future applications will not be confounded by
unraveling duplicative and/or mismatched phenotypic designations.
Equally important is to approach functional genomics, proteomics and metabolomics from an
integrative systems biology perspective. Within a systems biology approach, each type of

Page 13
biological information (DNA, RNA, protein, protein interactions, biomodules, cells, tissues,
etc.) also has individual elements (e.g., specific promoters, genes or proteins), and the
interrelationships of all these elements and types of biological information must be
determined and integrated to obtain a view of the system as a whole. What is ultimately
desired is the ability to unravel the complexities of epistatic (genotype by genotype; GXG)
and genotype by environment (GXE) interactions and how they affect phenotypic expression.
A typical GxE interaction of concern for agricultural production would be the change in
performance of a set of genotypes among differing environments. Deciphering these
complexities requires a holistic approach that describes and understands the biology
underlying phenotypes.
The post-genome sequencing era will bring enormous quantitative and phenotypic data to the
table. The USDA is the logical organization to lead this systems biology approach for
agricultural species. It is suggested that compartmentalization of genomics programs, as has
been done in the past for both CSREES and ARS program management, should be shifted
toward integration of functional genomics approaches into all program areas and disciplines
(e.g., animal growth and production, animal health, animal well-being, aquaculture, food
safety, animal waste management, animal and human nutrition, etc.). A cross-disciplinary
research effort will be required to integrate the global genomics data into information that is
usable and applicable across the diverse landscape of agricultural production.
Specific Conclusions and Recommendations from the Functional Genomics Module:
1) Downstream work in functional genomics and proteomics will be where the big
payoff from animal genomics research is reaped.
2) Develop a clear and standardized methodology for defining phenotypes for success,
particularly in the emerging areas of animal health and well-being.
3) Utilize a “big science”, “holistic” approach to unravel the complexity of epistatic
and genotype by environment interactions. [Agricultural animal genomics research is
ideally suited to an integrative systems biology approach]
4) Significantly enhance the bioinformatics capacity within the public agricultural
animal research enterprise to handle the increasing complexity and volume of
genomic and proteomic data.
5) Substantially increase the comprehensive funding for downstream functional
genomics, proteomics, metabolomics, and metagenomics research in agricultural
animal species to capitalize on the previous investments in genomic resources, tools,
and reagents.
6) End the previous separation of genomics efforts within USDA research portfolios as
the integration of functional genomics approaches as a foundation in all program
areas and disciplines is warranted.
7) To integrate genomic approaches across disciplines, improve the coordination and
effectiveness between ARS and CSREES by developing and implementing a long-
term strategic plan for USDA animal genomics research.
Focusing on Bioinformatics Resources. The third and final module of the workshop
focused on bioinformatics needs. Elliott Margulies of the NIH’s Intramural Sequencing

Page 14
Center discussed the vision of the post-sequencing era after sequencing the human genome
(Collins et al, 2003). Because the human genome is extraordinarily complex and its function
is poorly understood, the grand challenge for NIH is to catalogue, characterize, and
comprehend the entire set of functional elements encoded in the human genome. Embedded
within the complexity of the genome is the fact that only 1 to 2% of the DNA sequence
actually encodes proteins, and the full complement of protein-coding sequences still remains
to be established. Consequently, a major role for comparative sequence analysis will be the
identification of functionally important non-coding sequences. These sequences will be hard
to identify, as virtually no complementary datasets are available across various species to
assist with computational predictions. Nevertheless, methodologies for multi-species,
comparative sequence analysis relative to the human genome exist and can be used to gain
insight regarding species divergence, as well as substitution rates within coding or non-
coding regions under natural selection pressures over time.
John Keele, USDA/ARS U.S. Meat Animal Research Center, provided an overview of the
current bioinformatics infrastructure for agricultural animal genomics. The most useful
databases and bioinformatics resources included those at The Institute for Genomics
Research (TIGR), National Center for Biotechnology Information (NCBI), MS-Access or
MySQL software for local sequence and genotyping databases, and DNAStar (soon to be
replaced by SeqWeb). In addition, the Generic Model Organism Database (GMOD), funded
by the NIH and the ARS, was mentioned as a unique tool for genome database visualization,
curation, and ontology. Adequate databases and tools are available to manage and analyze
ESTs, SNPs, microarray, SAGE and proteomics information; however, there remain unique
personnel, skills, and software needs for each of these tools. It was noted that the general
lack of bioinformatics personnel and minimal integration of relational databases with all
aspects of research are the two critical factors that are limiting progress in the field of
Specific Recommendations from the Bioinformatics Module:
1) Focus USDA resources on its unique capabilities, such as phenotypic
characterization, population and quantitative genetics, physiology, etc., and be
careful to not “re-invent” the bioinformatics capabilities already in place in other
genomics research communities.
2) Immediately provide training programs and associated support for faculty
sabbaticals, postdoctoral associates, and graduate students focused on integrating
biology and computing, since one rate-limiting step for USDA in bioinformatics is
awareness and literacy in use of existing tools and lack of basic training programs to
bring new bioinformatics personnel online.
3) Develop standard descriptions of phenotypes as this is a second rate-limiting step for
USDA in bioinformatics and it is a problem that will be exacerbated when functional
genomics research moves into the more challenging areas of animal health and well-
being in the near future.
4) Create a USDA bioinformatics working group at the Research, Education, and
Economics mission area level to: a) coordinate and define ARS and CSREES efforts
among data producers, tool developers, analysts, and consumers; and b) to better
coordinate with other Federal and international agencies.

Page 15
5) To best serve the bioinformatics needs of agricultural animal genomics, leverage
USDA resources with others to develop expertise and new tools.
a. Support species-specific annotation;
b. Organize curation groups for management of livestock genome sequence
resources, in concert with existing groups (i.e,. NCBI, UCSC, Ensembl), to
help build browsers with characteristics important to current and future
animal genomics research.
c. Link genomic data to published literature in the animal sciences.
d. As large numbers of SNP are discovered and validated, develop databases
linking haplotypes with phenotypes and further tools (e.g., NCBI, dbSNP) to
facilitate QTL mapping and association studies for multiple species.
e. Develop a centralized and standardized system for microarray analysis and
gene expression databases by requiring agreement on database platform (s)
for microarray target annotation and gene expression data mining with
intentions to link to genome assemblies and associated gene and protein
There is little doubt that the investments made to date in animal genomics will yield
enormous dividends in the future for the producers and consumers of animal products and for
the biomedical sciences. However, this workshop clearly identified a number of areas that
need significant programmatic and funding attention within the USDA research infrastructure
for this potential to be realized in a timely manner. Opportunities appear to exist and should
be explored further for leveraging of future efforts with other Federal programs, given the
wealth of genotypic and phenotypic information catalogued on pedigreed agricultural animal
populations. Furthermore, there was strong consensus that in the post-sequencing era,
research employing genomics techniques and tools should be integrated across all disciplines
engaged in the animal sciences as opposed to being separated into “genomics” program
areas. An overwhelmingly clear message from the workshop was that it is critical for USDA
research leaders to develop and implement a visionary, long-term plan for animal genomics
research as soon as possible. Such a plan will ensure that the full potential of past, current,
and future efforts and investments in animal genomics will have a positive impact on animal
producers and the public in the post-sequencing era.
The authors gratefully acknowledge the scientific contributions of D. Adelson, L. Alexander,
S. Burgess, A. Capuco, H. Cheng, N. Cockett, L. Cogburn, E. Connor, P. Coussens, J.
Dodgson, C. Elsik, C. Ernst, K. Eversole, L. Gasbarre, J. Keele, H. Lewin, H. Lillehoj, E.
Marguiles, J. Ostell, J. Reecy, C. Rexroad, III, R.M. Roberts, G. Rohrer, M. Rothschild, D.
Schneider, L. Schook, T. Smith, T. Sonstegard, J. Vallet, C. Van Tassell, G. Waldbeiser, W.
Warren, J, Womack, K. Worley, and K. Zuelke. Additional appreciation is extended to P.
Brayton, P. Burfening, L. Ellis, D. Hamernik, J. Jen, P. Johnsen, S. Kappes, M. Mirando, J.

Page 16
Peterson, and W. Zamer for their participation in the planning and implementation of the
workshop. Finally, the efforts of the national program staff of CSREES, especially V.
Martin, and ARS, especially L. Mangra, for handling the logistics of the workshop are
sincerely appreciated.
Alexander D.J. (2002) Should we change the definition of avian influenza for eradication
purposes? Avian Diseases 47:976–81.
Andersson, L. et al. International Chicken Polymorphism Consortium (2004) A genetic
variation map for chicken with 2.8 million single-nucleotide polymorphisms. Nature
Andersson L. and Georges M. (2004) Domestic-animal genomics: deciphering the genetics
of complex traits. Nat Rev Genet. 5(3):202-12.
Casas, E., Keele, J.W., Shackelford, S.D., Koohmaraie, M., Sonstegard,,T.S., Smith, T.P.,
Kappes, S.M., and Stone R.T. (1998) Association of the muscle hypertrophy locus with
carcass traits in beef cattle. Journal of Animal Science 76(2):468-73.
Collins, F.S., Green, E.D., Guttmacher, A.E., and Guyer, M.S.. (2003) Nature 422:835-47.
Gibbs, R.A. et al. International HapMap Consortium (2005) A haplotype map of the human
genome. Nature 437:1299-1320.
Gibbs, R.A., Weinstock, G., Kappes, S.M., Schook, L.B., Skow, L., and Womack, J. (2002)
Bovine genomic sequencing initiative: De-humanizing the cattle genome.
Diederichs, S., Baumer, N., Schultz, N., Hamra, F.K., Schrader, M.G., Sandstede, M.L.,
Berdel, W.E., Serve, H., and Muller-Tidow, C. (2005) Expression patterns of mitotic and
meiotic cell cycle regulators in testicular cancer and development. International Journal of
Cancer Mar 30; [Epub ahead of print].
Frebourg, T., Oliveira, C., Hochain, P., Karam, R., Manouvrier, S., Graziadio, C., Vekemans,
M., Hartmann, A., Baert-Desurmont, S., Alexandre, C., Lejeune Dumoulin, S., Marroni, C.,
Martin, C., Castedo, S., Lovett, M., Winston, J., Machado, J.C., Attie, T., Jabs, E.W., Cai,
J., Pellerin, P., Triboulet, J.P., Scotte, M., Le Pessot, F., Hedouin, A., Carneiro, F., Blayau,
M., and Seruca, R. (2005) Cleft lip/palate and CDH1/E-cadherin mutations in families
with hereditary diffuse gastric cancer. Journal of Medical Genetics 14; [Epub ahead of

Page 17
Ghosh, F., Wong, F., Johansson, K., Bruun, A., and Petters, R.M. (2004) Transplantation of
full-thickness retina in the rhodopsin transgenic pig. Retina 24(1):98-109.
Gkoutos, G.V., Green, E.C., Mallon, A.M., Hancock, J.M., and Davidson, D. (2004)
Building mouse phenotype ontologies. Pac Symp Biocomput. 9:178-89
Grobet, L., Martin, L.J., Poncelet, D., Pirottin, D., Brouwers, B., Riquet, J., Schoeberlein, A.,
Dunner, S., Menissier, F., Massabanda, J., Fries, R., Hanset, R., and Georges, M.A. (1997)
Deletion in the bovine myostatin gene causes the double-muscled phenotype in cattle.
Nature Genetics 17:71-4.
Hamernik, D.L., Lewin, H.A., and Schook, L.B. (2003) Allerton III. Beyond livestock
genomics. Animal Biotechnology 14:77-82.
Hirano, T., Yamauchi, N., Sato, F., Soh, T., and Hattori, M.A. (2004) Evaluation of RNA
interference in developing porcine granulosa cells using fluorescence reporter genes. J
Reprod Dev. 50(5):599-603.
Jensen, L.R., Amende, M., Gurok, U., Moser, B., Gimmel, V., Tzschach, A., Janecke, A.R.,
Tariverdian, G., Chelly, J., Fryns, J.P., Van Esch, H., Kleefstra, T., Hamel, B., Moraine, C.,
Gecz, J., Turner, G., Reinhardt, R., Kalscheuer, V.M., Ropers, H.H., and Lenzner, S.
(2005) Mutations in the JARID1C gene, which is involved in transcriptional regulation and
chromatin remodeling, cause X-linked mental retardation. American Journal of Human
Genetics 76(2):227-36.
National Academy of Sciences. (2002) Exploring Horizons for Domestic Animal Genomics:
Workshop Summary. (Ed. Pool R and K Waddell). National Academy Press, Washington
DC (42 pp).
Nichols, K.E., Hom, J., Gong, S.Y., Ganguly, A., Ma, C.S., Cannons, J.L., Tangye, S.G.,
Schwartzberg, P.L., Koretzky, G.A., and Stein, P.L. (2005) Regulation of NKT cell
development by SAP, the protein defective in XLP. Nat Med. 11(3):340-345.
Paradis, F., Vigneault, C., Robert, C., and Sirard, M,A. (2005) RNA interference as a tool to
study gene function in bovine oocytes. Mol Reprod Dev. 70(2):111-21.
Petters, R.M., Alexander, C.A., Wells, K.D., Collins, E.B., Sommer, J.R., Blanton, M.R.,
Rojas, G., Hao, Y., Flowers, W.L., Banin, E., Cideciyan, A.V., Jacobson, S.G., and Wong,
F. (1997) Genetically engineered large animal model for studying cone photoreceptor
survival and degeneration in retinitis pigmentosa. Nature Biotechnology 15(10):965-70.
Roberts, R.M. (2001) The place of farm animal species in the new genomics world of
reproductive biology. Biology of Reproduction 64:409-17.
Schmidt, C.W. (2004) Metabolomics: what’s happening downstream of DNA.
Environmental Health Perspectives 112(7):A410-5.

Page 18
Venter, J.C., Remington, K., Heidelberg, J.F., Halpern, A.L., Rusch, D., and Eisen, J.A., et
al. (2004) Environmental genome shotgun sequencing of the Sargasso Sea. Science
Wilson, R. et al. Inernational Chicken Genome Consortium (2004) Sequence and
comparative analysis of the chicken genome provide unique perspectives on vertebrate
evolution. Nature 432:695-716.
Wall, R.J., Kerr, D.E., and Bondioli, K.R. (1997) Transgenic dairy cattle: genetic
engineering on a large scale. Journal of Dairy Science 80(9):2213-24.
Wall, R.J., Powell, A.M., Paape, M.J., Kerr, D.E., Bannerman, D.D., Pursel, V.G., Wells,
K.D., Talbot, N., and Hawk, H.W. (2005) Genetically enhanced cows resist
intramammary Staphylococcus aureus infection. Nature Biotechnology 23(4):445-51.
Wong, E. (2004) Poultry Research in the Post-Genome Era. ISB News Report.
Yamagishi, S., Nakamura, K., Inoue, H., and Takeuchi, M. (2005) Met72Thr
polymorphism of pigment epithelium-derived factor gene and susceptibility to age-related
macular degeneration. Med Hypotheses 64(6):1202-4.
Yang, J., Ratovitski, T., Brady, J.P., Solomon, M.B., Wells, K.D., and Wall, R.J. (2001)
Expression of myostatin pro domain results in muscular transgenic mice. Mol Reprod Dev.