About Katmandoo

Katmandoo is a biosciences data management system initially created for use in plant breeding and evaluation programs where the aim is selection of superior genotypes in terms of a range of traits (eg. grain yield, quality characteristics and disease resistance). It aims to provide a single tool for managing phenotypic and genotypic data at both an individual and multiple experiment level. The system has much wider application, however, being appropriate for data from other programs of designed experimental research within the biosciences.

The Katmandoo system consists of two (separate) fundamental components: a relational database to store the information and a software application to access it. The Katmandoo application is being developed as a suite of modules:

• trial data management (completed)
• genealogy (pedigree) management (beta)
• tools to capture trait score using Windows-Mobile OS (beta)
• crossing tool (v2.5)
• molecular marker subsystem (v2.5)
• tools to integrate phenotypic and genomic data (v3.0)
• seed inventory subsystem (v3.5)

where not all need be present depending on the level of functionality required.

Katmandoo is not an analysis system; statistical analysis and experimental design are areas of active research as is the specialist software necessary to implement them. A key feature of Katmandoo is the ability to import, store and export all information necessary to allow a rigorous statistical analysis for a wide range of designed experiments. The accommodation of multi-environment trial data means that Katmandoo is ideal for the ever-growing scenario of inter-organizational and national scale programs in both plant and animal research.

The flexibility of Katmandoo to meet emerging biological and statistical needs is a key property of the underlying data model. The storage of phenotypic data is made with reference to the distinction between the experimental unit and the observational unit (that is, the unit on which a measurement is made). The generality this allows is most easily illustrated by example; consider the following related to crop improvement programs:

• Plant variety field trial: In such trials, the treatments (varieties) are randomised to field plots using some sort of experimental design. Katmandoo stores data associated with both the experimental unit (field plot) and the observational unit, which in the case where one observation per plot is taken, are notionally the same. The information at the plot or experimental unit levels includes the variety assigned to it and design information such as replicate number, field column, field row and incomplete block etc. The data are stored at the observational unit level and may include response variables such as grain yield and covariates such as insect damage. If several  samples are taken per plot, such as dry matter quadrats, these constitute multiple observational units from the same experimental unit.

• Plant variety laboratory trial: In such trials, grain from field plots is sampled then processed in a laboratory to generate the trait of interest (eg. milling yield for wheat). These are known as multi-phase trials. The milling yield example has two phases, namely the field trial (phase 1) and the laboratory trial (phase 2). Replication and randomisation are employed in both phases of the design. Phase 1 data are stored as in the previous example but for phase 2 data the field plot essentially becomes a "treatment" in the laboratory and the experimental units correspond to grain samples taken from the plots. The design variables associated with the experimental unit would include field replicate, field column, field row, lab replicate, day of milling, time of day and mill operator. Katmandoo stores these as attributes (virtual columns) of the experimental unit while the data, in this case milling yield, is kept at the observational unit level. The measurement of some quality traits involves more than two phases (eg. dough extension traits require a field, milling and extension phase); Katmandoo stores all the necessary design factors (that is, from all phases) and the data (including any subsampling) at the appropriate levels.

• Series of plant variety trials: Varietal selection is usually based on information from a series of trials, also known as multi-environment trials (METs) in which similar sets of varieties are tested in a range of geographic locations and possibly over several seasons. Katmandoo stores the raw data, design and treatment information as above, summarised and meta information from each experiment to facilitate a joint analysis across experiments at the trial mean level if necessary, and the results of any such MET analysis from trial summaries or raw plot data.

• Known genetic relationships: Another goal of experiments such as those above is to select varieties for use as parents in the breeding program. Statistical analyses can be extended to incorporate pedigree information about the varieties from the genealogy stored in Katmandoo, which can be exported together with the phenotypic data for analysis. 

• Identification of Quantitative Trait Loci (QTL): Katmandoo uses special data structures to store molecular information on individuals. The marker, phenotypic and observed pedigree data are integrated and available for export.

Katmandoo has application beyond crop improvement programs. Data from any type of field trial may be stored, including data from factorial experiments (where the treatments comprise combinations of levels of several factors rather than levels of a single factor). Other examples include glass-house experiments and animal experiments, including experiments where repeated measurements and multiple traits are recorded on experimental units. 

The database component is built on a normalised data model implemented in SQL Server 2000 / SQL Server 2005 / SQL Server 2005 Express Edition and the Katmandoo application is developed in C# using the Microsoft .NET Framework 2.0.

Minimum recommended system configuration:

• Microsoft Windows XP or Windows 2000 SP4 or later
• .Net Framework 2.0 SP1 redistributable package^*
• Microsoft Data Access Component (MDAC) 2.8 or later^*
• Microsoft SQL Server 2005 Express Edition^* or SQL Server 2000 or SQL Server 2005 or MySQL 5.1 (or later)
• Microsoft Excel and Microsoft Word 97 or later

^* Provided by the installataion package

The Katmandoo database and application binary are distributed free of charge subject to the conditions in the Katmandoo license.

Please click here for contact information and responsibilities of the team members.

We would like to thank Chris Lisle¹, Damian Collins¹, David Luckett¹, Neil Coombes¹, Peter Martin¹, Gregory Lott⁴, David Jordan², Colin Cavanagh³, Marcus Newberry³ and our colleague at NSW DPI and QLD Primary Industries and Fisheries for their support during the system development of Katmandoo and KollecTrait.

¹NSW Department of Primary Industries
²QLD Primary Industries and Fisheries
³CSIRO Plant Industries, Canberra
⁴Molecular Plant Breeding CRC, University of Adelaide