Objective 2.2 Implementation Team Meeting 20101001
Minutes of the IBP Meeting held on Friday afternoon 1st October 2010, Tucson, Arizona
Chengzhi Liang, IRRI
Diego González-de-León, GCP
Fran Clarke, AAFC
Graham McLaren, GCP
Guoyou Ye, IRRI
Héctor Sánchez Villeda, CIMMYT
Jean-Marcel Ribaut, GCP
Oscar Riera-Lizarazu, ICRISAT
Peter Wenzl, CIMMYT
Trushar Shah, ICRISAT
The meeting was convened by Jean Marcel Ribaut to review the progress of the iPlant discussions, assess the impact of the outcome of the iPlant meeting on the broader IBP objectives and revise plans for other IBP objectives in the light of these outcomes.
Discussions with iPlant regarding collaboration on Objective 2.3 - The Development of the Integrated Breeding Workbench were very positive with iPlant and IBP recognizing the synergy of objectives in this area. Both iPlant Grand Challenges were positive about the collaboration and its benefit to them and it fits well with iPlants Discovery Environments.
We discussed the implications of this collaboration and agreed that CIMMYT, ICRISAT and IRRI would participate/lead, to various extents, in the collaboration to define modules, supply specifications, test prototypes and support the tools as they become available.
For both the IB Fieldbook and the Analytical R Pipeline modules, identified as top priorities during the meeting, basic programming during the first two iteration cycles of module development should be done at iPlant. IBP programmers will take on from that stage to refine module packaging based on testing and strong input from users. Integration of other modules into the iPlant Cyber Infrastructure to develop an integrated analytical workflow for modern breeding will be addressed in a stepwise manner. On a longer term, coding activity will move back to the participating centers as the Workbench evolves. This is the vision of iPlant - to have external collaborators using their infrastructure to imbed applications into their cyber-infrastructure.
From our meeting, it became quite clear that the iPlant collaboration, even with the IB Fieldbook given full priority, will not deliver a standalone IB Fieldbook (FB) the IBP needs in the short term. Through that collaboration we hope to achieve more efficiently our mid and long term objectives producing an integrated workflow system for modern breeding that will be integrated into a sustainable platform. Hence it was decided to continue with the original
plan of re-developing the CIMMYT Fieldbook in the form of a configurable version, which would be deployed for crops other than maize by early 2011.
It was agreed that the best strategy to achieve this re-development was to have Diego Gonzalez-de-Leon coordinate the specification of the revised fieldbook in terms of functionality and interface, and to have Hector Sanchez of CIMMYT lead a group of java developers to implement the application in rapid cycles of development; this would include iterated testing and feedback from CIMMYT's wheat and maize programs, as well as partners from ICRISAT and IRRI (a process facilitated by Diego). The goal is to rapidly develop a beta IB-FB version that covers both wheat and maize use cases and has an open, evolvable design so that it can be rapidly adapted to other crops. GCP will support the costs of this development. This arrangement has been endorsed by CIMMYT Management, and participants in the iPlant meeting from IRRI, ICRISAT and AAFC all agreed to be involved in the design and testing of the revised fieldbook.
Define the structure and functions of the IB Fieldbook, as well as the stepwise development stages that would form the early versions of the application; this will be sent to collaborators at the end of the first week of October.
Join forces to identify a couple of Java programmers to work with Hector on the coding (Programmers on board by Oct 15)
Mockup version of the fieldbook to be developed by Diego (first version Oct 22, final version after 3 cycles of iterations Nov 30)
Beta version of the FB ready for testing for maize and wheat (Jan 30)
Beta version tested by Centers during February and by users during the workshop (February 14-18) that will take place in IITA (use cases for cowpea, cassava, rice and sorghum will be represented).
Beta version implemented and adapted to several crops (maize and wheat, and 2-4 crops from the use cases) based on user feedback (April 31)
Deployment of the FB (ongoing process)
The mockup version will serve as reference for the development of both the Java standalone version described above and the prototype that would be integrated later in the iPlant platform. Therefore people trained to use the Java version should have no problem in adjusting to the module that will be accessible through the configurable workflow. It is important to underline that a web-based version of the Fieldbook could be installed as part of a virtual appliance
and used locally through any web-browser application in the absence of an internet connection.
Gouyou Ye indicated that IRRI wanted to develop a Java version of the "ICIS Light" Excel macro built at IRRI over the last few years, but he is willing to put this on hold and contribute to the testing of the beta version of the FB prototype in January.
The development of the IB Analytical Pipeline (basic statistical package based on R scripts) was also discussed - this, indeed, is the "second" first priority for the IB Workbench. The scenario here is different as the development of such a module should be more straightforward compared to the FB and a beta version of that prototype should be available by January 2011, with a first release in April. Diego is already compiling a list of candidate R scripts for inclusion in the IB Analytical Pipeline. It is important to emphasize that this Analystical Workbench will include many more analytical tools and options than those provided by the current Maize Fieldbook; indeed, most (and eventually all) analytical workflows in current use in the Centers will be progressively incorporated, including tools for genotypic and integrated analyses for which various R programs are under active development at CIMMYT and the University of Wageningen.
An informatic use case that we might develop first corresponds to the Additive Main effects and Multiplicative interaction analysis (AMMI), a statistical package based on R scripts. In addition, ICRISAT will also contribute scripts in R for basic statistical analyses (some of them are already embedded in the iMAS tool). Gouyou reported that IRRI might work on the development of an analysis package in R based directly on CropStat, with a menu interface (based on Tcl/Tk, a GUI scripting tool). We agreed that it would be suitable to join forces and Gouyou was keen to collaborate on the design of menu interfaces and selection of R scripts.