Introduction
The National Bureau of Plant Genetic Resources (NBPGR) is the nodal organization in India for the exchange, quarantine, collection, conservation, evaluation and systematic documentation of plant genetic resources. Since 1993, NBPGR has been implementing a germplasm registration system to document unique germplasm as well as to give credit to researchers who have (1) developed experimental material including parents and inbred lines or (2) identified promising germplasm (NBPGR, 2006). As an outcome of such an institutional mechanism to recognize the intellectual property inherent to the identification of elite genetic resources and consequently to promote their use, information on registered germplasm is published regularly in the Indian Journal of Plant Genetic Resources (Singh, Reference Singh2002) and special crop germplasm inventories are circulated (NBPGR, 2006; Kak et al., Reference Kak, Srinivasan and Sharma2009; Kak and Tyagi, Reference Kak and Tyagi2010). The primary objective of publicity is to promote the flow of unique germplasm to the ongoing breeding programmes.
The advent of information and communication technology revolution throughout the world has brought basic computational resources to each researcher in the form of either personal computers or shared computer laboratories. As a result, searching information in databases (wherever available) instead of printed registers and inventories has become the order of the day among students, researchers and even at remote experimental farms. Computer-assisted search identifies most suitable options quickly and easily from a vast data source, thereby tremendously promoting access to the desired information.
In order to relay the information on registered germplasm to the desktops of breeders and other interested clients, it is necessary to develop a simple and easy-to-use customized search engine. Search engines are widely used in today's computing world with various purposes, and they come in a range of forms, complexity, skill requirement, etc. (Grappone and Couzin, Reference Grappone and Couzin2011; Wikipedia (2012)). The nature of a search engine depends largely on (1) the type of data, (2) functionalities of the system and (3) hardware and software sophistication, and hence search engines end up as closed systems that may not always be universally compatible and portable. Search engines can be general (Google, Yahoo!) or subject specific (Entrez, AGRIS). Search engines can be limited (desktop search engines) or extensive (meta-search engines). The superior the search engine in its architecture and sophistication, the greater the requirement of computational power and Internet connectivity (Kreinovich, Reference Kreinovich2012). On the contrary, typically a registered crop germplasm search engine should serve both clients with adequate Internet access and contemporary computational resources and clients without a reliable Internet facility and having basic computer facilities.
We have developed a search engine solution that requires no installation procedures, needs no Internet connectivity, runs from a compact disk (CD) or a pen drive, demands nothing from the computer except for a default Internet browser (e.g. Internet Explorer), provides intuitive search facility and requires no specialized experience to use. We report here the employment of such a solution to develop a portable search engine for the Inventory of Registered Crop Germplasm of NBPGR.
Materials and methods
Technologies involved
Data were transformed into an eXtensible Markup Language (XML) format, an industry standard for handling a large amount of data in text form without a database management system (Fung, Reference Fung2000; Chaudhri et al., Reference Chaudhri, Rashid and Zicari2003) and sharing both the format and the data on the Internet. JavaScript, an open-source programming language, was implemented to enable a standalone or Web-based solution to integrate complex data-related functionality (Bayross, Reference Bayross2005) by interacting with Hyper Text Markup Language (HTML) source code. Decoding of the XML-tagged data by the JavaScript language into an application was carried out by an XML parser. The application output was encapsulated in HTML pages and presented via Web browsers in online and offline modes (Fig. 1).
Fig. 1 Data flow diagram of the portable search engine. A colour version of this figure can be found online at http://journals.cambridge.org/pgr.
Architecture of the search engine
The architecture of the portable search engine consists of four distinct units (Fig. 2), which are introduced below.
Fig. 2 Architecture of the portable search engine for registered crop germplasm. A colour version of this figure can be found online at http://journals.cambridge.org/pgr
Loading XML data application (block 1)
First, the unit of the architecture links and then loads the XML-based data source (local or Web based) to the search application. The module ensures the loading of the XML data into the computer memory before the search application operates.
Search input (block 2)
Searching the database is initiated and loaded via a Web browser. A search box-based query initiates the search process. Currently, the minimum number of search characters to be entered is 3.
Running the algorithm (block 3)
This module accepts the user input with a Boolean logic. The algorithm searches sequentially in every record of the XML data file and transfers all matches to a temporary storage. The algorithm iterates the process until the end of the file.
Results display (block 4)
Results are displayed as a formatted HTML page. The results are saved by copy–pasting into applications such as Microsoft (MS)-Excel, MS-Word, plain text file in comma separated values (CSV) format, etc.
User interface
User interface of the search engine opens in the default Web browser of the computer (fully compatible with old and new versions of all major Web browsers). A single-line text box is provided to accept the user's query input. The search box supports single or multiple words search term(s) with the Boolean operator AND to narrow down the search.
Result page
Hits produced by the search engine are presented in a tabulated format. If a search term, fully or partially, matches with any of the records, then the entire accession record is moved to the result pools that are arranged row-wise. Each result row contains the information on crop name, botanical name, national identity (accession number), donor identity, INdian Germplasm Registration (INGR) number (registration number), year, pedigree, developer, developing institute and novel unique features. The result page also displays the search term, the number of matched records and the total number of records available in the IRCG database.
Results
The portable search tool was implemented to query the registered germplasm data. Since 30 December 2011, as many as 1030 genotypes possessing unique traits have been registered with NBPGR (Tyagi and Kak, Reference Tyagi and Kak2012). When new records are available, the application would make a provision for users to update the data file by adding new rows. This provision helps to keep the application updated. Users, however, cannot incorporate additional information in the existing records.
Search engine
The primary objective of developing a portable search application is to cater to resource-challenged end users. The portable application carries (1) data, (2) portable search engine and (3) Web browser as the interface.
The algorithm used in the application is designed to consume minimum system resources and respond back in minimum time without affecting the system performance. Our tests showed that the application could handle up to 100K records without perceivable differences. For instance, a search with two AND operators (i.e. three search terms) utilized 2.7 megabyte (MB) memory and 5 ms time for application loading, and 14.7 MB memory and 7.3 s time for scripting and rendering. The observations confirm that the application is built for searching the XML data file with minimum memory and time overheads. This feature makes the application compatible with even old computers and browser versions.
Search terms for registered crop germplasm
The user is provided with a free-text search box and can search literally for any word (at least three letters). However, structured search is targeted against the contents of germplasm registration data, which are maintained for ten fields, namely crop name, botanical name, national identity (accession number), donor identity, INGR number (registration number), year, pedigree, developer, developing institute and novel unique features. Hits can be narrowed or expanded by using AND or OR between multiple search terms, respectively. They can be restricted to only exactly matching ones by using a space before or after the search term.
Search methodology
It is better to begin the search with a term matching any of the ten fields. For instance, to know whether there is any wheat genotype registered, the search term ‘wheat’ is used, which returns 152 hits. Alternatively, specifying wheat registered in 2008 (‘wheat and 2008’) returns 14 hits. Further narrowing the search with the location of the germplasm (‘wheat and 2008 and flowerdale’) returns three hits. Searching for durum wheat genotypes registered in 2008 (‘wheat and 2008 and durum’) results in a solitary hit. Searching for trait-specific tomato germplasm (‘tomato and high total soluble solids (TSS)’) returns a single hit. The search engine is provided with an illustrative help file (http://www.nbpgr.ernet.in/IRCG/html/IRCG_EXAMPLE.pdf).
Discussion
Search engines are not new in agricultural research; they have been used in several applications as an essential tool. Many genebanks and international organizations have databases to facilitate germplasm searches in a comprehensive manner. Some of these Web portals are hosted by Genesys (www.genesys-pgr.org/), the National Institute of Agrobiological Sciences of Japan (www.gene.affrc.go.jp/databases-plant_search_en.php), the International Rice Research Institute (www.iris.irri.org/germplasm/), the Centre for Genetic Resources of Wageningen (www.cgn.wur.nl/applications/cgngenis/), the National Plant Germplasm System of the USA (www.ars-grin.gov/npgs/searchgrin.html), etc. Major genebanks around the world have implemented robust databases and Web portals not only to facilitate access to information but also to act as a comprehensive genebank management system. Even a cursory browsing of these excellent online resources shows that using these germplasm information systems requires good Internet connectivity. Further, most of these portals employ latest tools of visualization and data presentation, and their best use is possible only with adequate computer resource replete with plug-ins and apps.
In order to realize the enhanced utilization of germplasm, it is imperative that information on germplasm is accessible to field researchers. In reality though, the absence of adequate infrastructure appears to thwart the connection between excellent databases and field workers devoid of requisite IT facility. To overcome this impediment, many organizations including the NBPGR produced and distributed CDs containing published data on genetic resources. However, (1) the methodology always lacked the provision to ‘search’ the data and (2) gaps appeared between the information available online and through CDs. We conceived a lightweight specialized search tool and the concept has been implemented in the form of a portable search engine to access unique crop germplasm registered at the NBPGR. The portable search engine along with data in the form of CDs was distributed among various plant breeders by the NBPGR. When new records are available, the users will update the excel file by adding new rows. This provision helps to keep the application updated. Users, however, cannot incorporate additional information in the existing records.
Conclusion
We present here a concept to develop a simple yet effective search tool as an offline solution to cater to people and places with low-end computational resources. The concept is implemented as a search engine for the NBPGR's registered crop germplasm. The new tool is expected to attract attention of breeders and other plant researchers, and is projected to facilitate search and eventually enhance the utilization of plant genetic resources.
Acknowledgements
The authors acknowledge the Director, NBPGR for facilities, Dr R. K. Tyagi for support and encouragement, and Mr Rajeev Gambhir and Mr Vijay Mandal for technical support. Vikas Kumar was financially supported by the Department of Biotechnology, New Delhi in the form of research associateship.
