“Elaborate statistics are not substitute
for meticulous experimentation.”
G.W. Snedecor

Trial network

7min

R&D IT tools to help qualifying and caracterizing a trial network

   

    

Click here to ask the video of the webinar "Field Trial Network Qualification"

Climate changes and diversification of environements imply increasing expression of biotic & abiotic stresses. Comparing varieties in this context is a great challenge. Therefore breeding objectives should consider not only the intrinsec agronomic value but also the ability of varieties to maintain their performance over various environments.

Breeders and testers want to study more than average values, they need to study the variability of the responses expressed in the contrasted trial network, with an evaluation of the "genetic by environment" (GxE) interactions. Efficient methods and tools provide this information, including standardized data collection, dynamic graphs to visualize and analyze the response of genotypes to environmental data.

To caracterize the trial network and optimize the interpretation of variety trials, mainly 3 tools can be proposed to the researcher:

Map representation of environmental factors collected on the trial networks that can be compared to variety performances.

Matrix and PCA analysis enabling to caracterize varieties based on GxE interactions

Clustering analysis enabling to group location of the trialo network based on agronomical diagnostic.

How to define a trial network within RnDExp®

   

More info at 5'19'' on the webinar replay "Field Trial Network Qualification"

In the trial network window, we have a view of the trial network, with a Multi-location Experiment (MLE) presenting the varietal list to dispatched among various locations named Single-Location Experiment (SLE). Below are displayed the modalities that will be studied over every location.

Every location is defined by its country, region and culture place, that can be geo-referenced. Then the experimenter can choose how many replications, which ones will be controls, then choose the material to evaluate and select the variables

Capture trial network heterogenenous data within RnDExp® (Environmental data, phenotypic observations)

   

More info at 11'55'' of the webinar replay "Field Trial Network Qualification"

Metadata on every experiment of the trial network can be captured in a standardized way in the "Metadata form":

 - Experimental  characteristics: plot surface, number of rows, orientation

 - Sowing information: date, measurements

 - Technical itinerary and human responsibilities

 - Scoring information

 - Location information such as weather, soil, topography, etc. can be added into the enalys for trial network characterization.

   

Collecting phenotyping data from the varieties in the fields can be imported by an Excel file or directly input in the breeding software.

It can be done also with data collection software such as RnDExp® Mobile

 

Trial network qualification thanks to visualizing and analyzing result data, enable to qualify the trial and take decisons

   

More info at 16'45'' of the webinar replay "Field Trial Network Qualification"

Web mapping representation

Directly connected with the centralized database, the web mapping application RnDExp Web® takes information from trial network metadata and phenotypic data already entered into the system.

Accessible from any browser on the web, it diplays heterogeneous information about the trial network, such as the measured pressure of diseases by experiment, the variations of pressure of disease by region, weather values, soil quality and any trait present in the RnDExp® database of the plant breeding software.

   

More info at 20'30'' of the Webinar replay "Field Trial Network Qualification"

Grouping locations of the trial network based on agronomical results

To qualify the various locations, researchers can group them depending on their homogeneity or differences. This report is based on the phenotype observations of the varieties.

Such information can be visualized in RnDExp for a vast range of traits:

  • yield results
  • Temperature levels
  • Nitrogene index
  • Disease pressure

The data is grouped thanks to cluster analysis on a selection of locations and  variables of of the field network. The cluster clearly shows groups of locations that qualify the trial network.

   

   

   

More info at 23'00'' of the webinar replay "Field Trial Network Qualification"

Caracterizing varieties of the trial network based on GxE interactions

Genetic by environment (GxE) interactions enable to qualify the trial network as well.

  • GxE Matrix: For one year and a list of locations, a GxE matrix can be genarated. In the example pictured here, the raws of the matrix represent varieties, displayed with genotype name and with the corresponding yield result (Hybrid mean). Columns represent locations in the trial network, with the corresponding yield mean performance. Every variety and every location can be compared separately by GxE standard deviation. Moreover, in the table, the specific values of GxE are displayed in figure and in colors, enabling to understand which locations are more suitable for a given variety, and which varieties have the best performance in one location of the trial network.
  • Principal Component Analysis graph: This interactive PCA graph represents varieties distributed in a 2D graph according to responses of varieties to variation of location. In the image you can see the user has used a lasso to select all the varieties on the top left of the graph (green dots). Hence these varieties are selected in the list on the left. The same analysis can be launched the other way, displaying the distribution of locations of the trial network according to the response of locations in the locations.
  • PCA graphs in a report: This static report displays for every location a distribution of all the varieties of the trial network sorted by their response to the location environmental conditions. This enables to compare the responses of the various biotic and abiotioc stresses of this location. Then, a dendrogram displays all the hybrids grouped by response to the location conditions. The third graph represents all the factors contributing to response variation between varieties and locations.

Conclusion: Qualify a trial network

   

Several tools within the breeding software enable to study the varieties and locations, and qualify the trial network. These tools assist the breeder in choosing the optimal variety depending on its performance over various locations, and as well according to its ability to give good results over various locations. Since this latter objective is so important for marketing a variety nowadays, GxE interactions are at the center of breeders interest, and Doriane is more and more involved in this topic with its customers. 

To go further in the analysis of such an important topic, Doriane has also conducted research on this topic, which have been published in two posters in 2015 and 2016, which you can read among our scientific publications.

Was this article useful for you ? (Anonymous) SEND
Leave a comment

Doriane, your R&D IT partner

Doriane Logo

About Doriane

The expertise of Doriane brings the three players of R&D the resources to organize, share and valorize their research projects:
RESEARCH DEPARTMENTS: Fast implementation of business processes ; Liberation of researchers creativity
IT DEPARTMENTS: Real-time centralization of multilocation data ; Security and integrity of data
MANAGEMENT: Maximization of research investments ; Research departments monitoring ; Intellectual assets valorization and sustainability

Contact

31 avenue Jean Médecin
06000 Nice, France
contact@doriane.com
+33 (0) 492 478 444
Customer Area

You'll receive a mail with your new password.

Technical Support
Attend a demonstration
TeamViewer client for remote demonstration

After downloading and executing the file contact Doriane for login and password.

Support remote control
TeamViewer client for technical support

After downloading and executing the file please send login and password to Doriane.