The Databases of the USDA/ARS Plant Genome Research Program

Stephen R. Heller
Beltsville, MD 20705 USA

Douglas W. Bigwood
Department of Plant Biology
University of Maryland
College Park, MD 20742 USA


This paper describes the creation and operation of a large domestic and international collaborative project in the area of plant genomics. Examples of how the system is made available via the World Wide Web are provided.


Food is a basic need of mankind. With both a growing population and increasing standards of living throughout the developed and developing countries of the world, the need for increased supplies of food is apparent and will grow considerably in the decades ahead. Biotechnology offers new tools for the agricultural scientist to help meet these needs. Biotechnology can help address and solve problems of water supply, water quality, chemical pesticide reduction, improved quality food products, biological controls of pests, and other related issues.

In 1991 the United States Congress appropriated funds for the start of the USDA Plant Genome Research Program, under the direction Dr. Jerome (Jerry) Miksche, to develop the necessary tools to meet the food needs of the 21st century. The program consists of two basic areas. The first area is for agricultural research efforts to map (and eventually, in some cases to sequence) genes of agronomic importance, characterize these genes, and develop new plants with improved traits of economic importance. The second area is to develop the necessary infrastructure and support to assist the latter area. The system needed to support the handling and manipulation of all this data is the heart of this project. Only by having quick, easy, and accurate access to the vast amount of scientific data generated from the research results of agricultural scientists around the world can one expect to need the food needs in the next century.

From the start of the project it was recognized that a large and sophisticated computer system would be needed. Also, it was clear that the project needed to be global in scale, since food is a universal need, and US agriculture contributes significantly to the US economy and sales of US agriculture products are one of the few items for which there is a positive trade surplus. Thus, the database and computer system effort was initiated at the onset, along with the scientific research. Cooperation and collaboration on a large scale, with diverse groups of scientists, virtually none of whom work for the same organization, is a difficult matter. However, one of us (SRH) had experience developing a similar multi-organizational project, the NIH/EPA Chemical Information System (1,2) and the experience from this project was used as the basis for developing the computer system for the Plant Genome program.

The Plant Genome Database (PGD)

The first step in the database effort for this project was to find the appropriate organization with the USDA to house the database. With an eye on the future of what libraries will be in the next century, it was decided to locate the database activities at the US National Agriculture Library (NAL), one of three (3) national libraries in the USA. This decision agreement was easily made owing to the very positive attitude of the senior management of NAL towards this project. The Department of Plant Biology at the University of Maryland was awarded a contract to manage the project and develop the system.

Within months of the initiation of the project the authors were hired to begin the database and computer system development. One of us (SRH) was given responsibility for the overall management of the database, coordination of the researchers funded by USDA, and not funded by USDA (our collaborators). The other (DWB) was given the responsibility to manage a team of programmers to develop and make operational the actual computer systems. The project was both helped and hindered by the revolution in computer networking during the formative stages of the project. When the project first started the Internet was in its infancy and the World Wide Web (WWW) did not exist. As these technologies began to develop and emerge it made many of our efforts outdated very quickly. A more detailed history of the database development is available elsewhere (4). Details about the main software used in the Plant Genome project, ACEBD, is also available elsewhere (5,6) The focus of this paper will be on the current WWW implementation of the database.


The system, which is now operational at the National Agriculture Library (NAL) and available at no cost to the world-wide agriculture genome community, is called the Agricultural Genome Information Server (AGIS). The home page of the system which has been developed can be seen on the next page:

Agricultural Genome Information Server

A Point Top 5% Site and A Magellan 4 Star Site

Welcome to the Agricultural Genome Information Server sponsored by the U.S. Department of Agriculture, Agricultural Research Service. AGIS is a cooperative effort between the University of Maryland, College Park, Department of Plant Biology and the National Agricultural Library. Browse and search the National Genetic Resources Program genome databases and related biological information.

What's New? (30 October 1996) | Help | Search AGIS Docume nt Collection | How to... | Resources and Credits | Services ]



Conferences and Meetings


Other Data Access Methods

Other Information Servers

Maintained by the Genome Informatics Group at the National Agricultural Library. Send questions and comments to: / 30 October 1996

Figure 1. Plant Genome WWW AGIS Home Page

As seen in Figure 1, there are a number of items available in the Plant Genome Database system. The first are the Databases, then there are Documents, listings of Conferences and Meetings of interest to the project, Tools, and lastly, other methods of accessing and obtaining data and information from the projects - namely by ftp and gopher. We will go through the major items and show examples of each in order to present the reader with a basic understanding of the size, scope, and amount of information available in the overall system.

In the WWW version of this paper (7) one can just click on the appropriate hypertext link and go to the desired information. Thus, by clicking on the word "Genome" (found directly under the heading : Databases) one goes to next page, shown in Figure 2, which shows all the databases currently in the system.

[ What's New? (30 October 1996) | Help | Search AGIS Docum ent Collection | How to... | Resources and Credits | Services ]

Search Databases

Plant Genome [ query | about ]

Livestock Animal Genome [ query | about ]

Other Organisms Genome [ query | about ]

[AGIS home page]

[Plant Genome | Animal Genome | Other Genome | Plant Reference | Insect Reference]

Genome Informatics Group / 30 October 1996

Figure 2. Plant Genome Databases and Documents

By clicking on the words "browse", "query", or "about" one then is able to either browse or search the database for particular information about the species chosen. By clicking on the word "about" one is able to learn more about the database, such as the name of the curator, where the database is being developed and maintained, and so on, before proceeding on to look at the actual database. The browsing capability allows one to look at entries in the database for the various classes or fields which are present. The number of entries for each class or field is also given in the main listing.

For example, in the case of RiceGenes, there are 32 data classes from which one can chose. They are shown in Figure 3 (October 1996 version of the system), one simply highlights the item of interest, clicks on it, and the computer system goes directly that information.

Allele (579)
Author (2895)
Chrom_Block (137)
Colleague (260)
DNA (2154)
Definitions (5)
Gene_product (174)
Germplasm (377)
Homoeology (69)
Image (287)
Isolate (7)
Journal (73)
Keyword (2966)
Locus (3314)
LongText (23)
Map (100)
Map_Data (10)
Method (3)
Motif (339)
MultiMap (39)
Paper (1466)
Pathology (36)
Polymorphism (2376)
Probe (2403)
QTL (42)
Sequence (2780)
Species (34)
Text (1083)
Trait (1)
Var_release (23)
View (5)
gMap (14)

Figure 3. List of Classes or Fields of information in the Ricegenes database.

As an example, we have selected QTL as the data class and then retrieved a particular data object: QTL qBlast-1-1. The result of this browsing is shown in Figure 4. The "view graphic" allows one to click here and go directly to a plot of the chromosome and the (labeled) markers on the chromosome (as can be seen in Figure 5).


QTL : qBlast-1-1

[view graphic]

Position Map Rice-1 Ends Left  0  
                         Right 40 
Positive Contains RG140 
Significant_loci RG612 
Trait Blast-1 
Location Greenhouse, Philippines 
LOD_score 11.9 
%_Variation_explained 32.5 
Allele_effect 0.38 
QTL_study QTL-Blast-CO39/Moroberekan 
QTL_Remarks QTL for blast lesion number at LOD 6.0.    
              Markers RG140 and RG612 define the most  
              probable region.  The most significant   
              marker was RG612, and statistics given   
              are for this marker only.  Allele_effect 
              refers to the mean difference in lesion  
              number between the two genotypic groups  
              carrying the CO39 (susceptible) and      
              Moroberekan (resistant) alleles at the   
              significant_loci marker.  Note that this 
              marker does not appear on the            
              Rice-BS125/2/BS125/WLO2 map, but is the  
              last marker on chromosome 1 on the map   
              produced by this QTL study.              

Figure 4. Details of the RiceGenes QTL entry for qBlast-1-1.

Now, by clicking on "view graphic", one gets the actual map, as shown below in Figure 5.

Figure 5. Graphic view of qBlast-1-1 QTL.

The graphics displays (which generate genetic maps, physical maps, and sequences) are generated on-the-fly as GIF images. They are scrollable, click-able, and zoom-able. If you click on the name of a data object (as was shown in Figure 6), then the text of that data will be retrieved. One drawback to the current display methods is that a new GIF must be generated for each action. A new interface, written in the Java programming language, will provide the user with a much more interactive environment.

To look at details of the trait information in Figure 4, one would click on the word "Blast-1" to the right of the word Trait and the following (Figure 6) appears:

Rice Genes

Trait : Blast-1

Name Blast resistance 
Evaluated_in QTL-Blast-CO39/Moroberekan 
Description This study evaluated both qualitative      
              and quantitative resistance to rice      
              blast disease.  For qualitative          
              resistance, the disease reaction in the  
              greenhouse was scored using a modified   
              scoring system based on the 0-5 scale.   
              An additional score of 3+ was added      
              between scores 3 and 4.  Round lesions   
              of about 1-2 mm with gray centers and    
              brown margins and capable of sporulation 
              were classified as 3.  Those with        
              round/elliptical lesions of about 2-3 mm 
              with gray centers and brown margins and  
              capable of sporulation were scores as    
            Three individual parameters of partial     
              resistance were measured during the      
              greenhouse phase of this study: diseased 
              leaf area, number of lesions, and size   
              of lesions.  These 3 traits were         
              evaluated on 4 replications of 131 lines 
              using the blast isolate PO6-6, and were  
              used to identify QTL associated with     
              partial resistance to blast.  These 3    
              components were significantly            
              correlated, and most of the markers      
              identified in the study affected all 3   
              parameters of blast.  For field tests, a 
              randomized complete block design was     
              used.  Two rows of a resistant cultivar  
              separated test entries, and 3 rows of    
              susceptible cultivars were planted       
              around the blocks.  The percentage of    
              diseased leaf area was evaluated         
              visually at 3, 4, 5 and 6 weeks after    
              sowing.  In Cavinti, two trials with two 
              replications were done.  In Sitiung, one 
              trial with two replications was done.    
            The Allele_effect values given for the     
              QTLs refer to the mean differences in    
              either lesion number or percentage of    
              diseased leaf area between the two       
              genotypic groups carrying the CO39 and   
              the Moroberekan alleles at the           
              significant locus marker.  This value    
              may be followed by links to specific     
              allele records.  In this case, the first 
              allele is the one which increased the    
              trait of blast resistance, the second    
              allele decreased blast resistance.       

Figure 6. Blast-1 trait details

While this short presentation does has not allowed for a full description of the many pieces of information in the different plant map databases, it is hoped this material has shown some of the content and capabilities of the system. With the database system being accessed over 200,000 times per month, it is clear that it is a very useful system. The full scope and impact of the system can only be assess by visiting and exploring the system (8).


The authors would like to thank Dr. Jerry Miksche for his help and encouragement in developing the database and his leadership in the project. Without him this database would still be only a plan, not a true living and used system.


(1) S. R. Heller, G. W. A. Milne, and R. J. Feldmann, "A Computer Based Chemical Information System", Science, 195, 253-259(1977).

(2) G. W. A. Milne, R. Potenzone Jr., and S. R. Heller, "Environmental Uses of the NIH-EPA Chemical Information System", Science, 215, 371-375(1982).

(3) The other two are the Library of Congress and the National Library of Medicine.

(4) "USDA Plant Genome Research Program, Advances in Agronomy, Volume 55, 113-166 (1995). In particular, see pages 147-154.

(5) J. M. Cherry, and S. W. Cartinhour, ACEDB, a tool for biological information. In Automated DNA Sequencing and Analysis, Edited by M. Adams, C. Fields, and C. Venter, 347-356 (1994). Academic Press: San Diego, CA.

(6) R. Durbin and J. T. Mieg (1991-). A C. elegansDatabase. Documentation, code and data available from anonymous FTP servers at, and

(7) A WWW version of this paper is available by pointing your browser to the url:

(8) The AGIS system is available by pointing your browser to the url: