Implementation of the Integrated Breeding Workflow System (IBWS) in a large breeding program
The IBWS comprises one or more crop databases connected via the IB Workbench to a number of informatics tools arranged in various workflow configurations. These elements are now ready to be deployed in small and large breeding programs.
We consider some of the technical steps required to implement the system in any breeding program, but the management of the conversion from an existing system is the most critical part of the implementation and for a large program this must be planned and executed systematically.
Management of the implementation of the IBWS in a large breeding program
The first critical element is the program management decision to implement the system. Once the management has decided to implement the system this decision must be conveyed in absolutely unambiguous terms to the whole breeding team. If anyone in the line senses hesitation at the management level the implementation will fail (as with all institutional change).
Secondly, management resources must be allocated to the implementation task. A senior person who knows the program, is committed to the change, and has an aptitude for information management must be assigned to the task of driving the implementation and forming the link between breeders, technicians and system developers. The amount of management resources available for this Application Manager position is the most important criterion for deciding on the pace of implementation.
Thirdly, review the infrastructure available for the Information System and decide on the overall configuration of the system. Where will the central crop database be housed? Do all breeding projects have adequate access to that system or will it have to be mirrored to different locations? How many breeding teams are involved? How often do they need to see each others' data? Can each team operate on an independent local database, or do several teams need to have access to current cycle data from other teams? What should be the frequency of loading data to the project central database so that all teams it? What is the strategy for global publication of programme information?
Finally, the Application Manager needs to work with the breeding teams to draw up an implementation plan. Will all breeding teams adopt at the same time? Will there be a staggered implementation? Which teams are early adopters? What is the training and support capacity for each breeding team?
Implementation at the Technical Level
The following steps need to be taken to get an IBWS up and running for any crop. If there is already a Central Database for your crop most of the steps in A, B and C may already be done, in any case you will need to work with the administrator of that central database to ensure compatibility with your local installation.
Firstly, identify a database administrator and a data curator and data manager(s). These can all be the same person if you want, but they need skills in database administration, and understanding of database designs, an understanding of the crop, the breeding systems, and the data, as well as skills in managing and manipulating data with Excel and MYSQL. This 'team' should familiarize themselves with IBWS, and the best resource for that is this IBP Portal. The empty database and software can be downloaded from the IBP Portal. Once this is done the following steps can be taken:
A. Set up the Germplasm Management System
- Review historical pedigree information and decide how well it has been recorded and hence how easy it will be to parse into the IBP genealogy management system. This is a data management task (not a user task) and the data manager needs lots of help from breeders and germplasm people to get it done right. Decide how much historical data is worth the effort of converting. Important sets are germplasm collection catalogues and international testing nurseries.
- Review the list of germplasm development methods in the database and check that all the ones in use for the current crop are available. If not, define them. (It may be useful to delete ones which are never used for the crop to reduce confusion, but never re-use identifiers).
- Compile a list of germplasm collection and development locations and load these into the location management tables of the database.
- Decide on naming conventions for germplasm, germplasm lists, nurseries and trials.
- Develop scripts and procedures to convert historical pedigrees to the IBDB.
B. Set up the Data Management System
- Review the existing trait dictionary (available on the IBP Portal) for the crop or make a trait dictionary if one does not exist. This requires reviewing all traits usually recorded for the crop and specifying some detail about how those traits are related, described and measured. A template has been developed for this purpose and is on the . This dictionary does not have to be complete as new traits can always be added, but it helps to get a good basis.
- Review historical evaluation data and decide how much is important to be loaded.
- Compile a list of germplasm evaluation locations and load these into the location management tables.
- Decide on naming conventions for trials.
- Annotate historical evaluation data with the Trait Identifiers and the Germplasm Identifiers, Location Identifiers using the IBP Phenotyping Template so that these data can be directly loaded into the database.
C. Set up the Inventory Management System
- Add seed storage locations to the location management system
- Check that the seed inventory properties in the Trait management System are appropriate and that the units are valid for the crop.
D. Set up local data management systems for breeding teams or genetic resources groups.
- Identify research groups who need to use the information management system. Decide whether they have the capacity to run the system themselves or need to have a central unit or an application manger run the system and deliver them lists and fieldbooks in a conventional way.
- For each team set up a local database connected to the central which has been established. This process is facilitated by the process of defining a project in the IB Workbench.
- Confirm that all breeding methods, locations, traits, and naming conventions are suitable for the team, else add them.
- Make sure that the group's current germplasm sets and nurseries are in the local or central system, so that all current germplasm has identifiers. This uses similar methods to section A above and needs to be done by a data manager.
- Train the users of the system in the techniques they need to use. This is best done with a general overview followed by what might be called 'context sensitive training'. That is, when they need to do a task - prepare a nursery, prepare a stock list, generate a fieldbook etc they can call on someone to come and show them how to do it, and then do it themselves the next time with less help. It is not always possible to do it this way and some training material is available for adaptation to different crops.