Molecular signature predicting plant properties
Modern measuring techniques yield an enormous amount of data about plants. But how do you know which are relevant for a certain property? Scientists of Plant Research International are ranking the data, remove errors, and correlate them with a specific plant property.
The taste of sweet pepper, the quality of starch in potatoes, or the discolouration of crisps. These are all plant properties that breeders would like to modify. But then they need to know how this can be done.
Our scientists are providing the building blocks for this. They are collecting all sorts of possible data about the plant at micro level: which genes, proteins and metabolites are found in a plant and which genes come to expression. And they are also measuring many characteristics of the plant during growth, on the field or in the greenhouse.
Correlating data
This yields an enormous amount of measurements of various plants of which the correlation is not immediately clear: the amount of data is simply too large and another fact is that there is a lot of variation between plants. The art then is to distinguish true biological effects from the more coincidental differences and the measuring errors, and to select those data that are relevant for the taste of sweet pepper or the discolouration of crisps.
The scientists are ranking the observations in databases by using computer programmes and statistical models to find correlations. And they are plotting data in graphs, figures and tables to visualise correlations. They are developing new tools for proper interlinking of the data. This yields a complete picture of how genes, proteins and metabolites are collaborating in networks and are thus determining the taste of sweet pepper or the discolouration of crisps.
This information enables breeders to breed varieties with better properties.
Integration of data from various measuring techniques with a property that is relevant for the breeder (flesh colour potato, taste sweet pepper, etc.) to identify the most important genes, metabolites and proteins that are involved in such a property. This is followed by network analysis to visualise the results.