- Version
- Download 4
- File Size 4.00 KB
- File Count 1
- Create Date April 21, 2023
- Last Updated April 21, 2023
D4.3 Predictive models for agro-biotechnology - Executive Summary
GLOMICAVE aims to exploit the information hidden in the existing scientific literature and large-scale omics datasets for a better understanding of genotype-phenotype relationships. In this deliverable, we will focus on the agro-biotechnology sector (Task 6.3 Fruit quality management and 6.4 Plant growth) and will present predictive models for those two cases, respectively the prediction of different fruit phenotypes using both metabolomics and transcriptomics data, and the prediction of cassava root developmental stages from transcriptomics data.
Fruit quality remains crucial to human nutrition and health and is an essential parameter in determining shelf life and purchase choice of the consumers. Fruit quality is significantly affected by agricultural practices prior to harvest, especially the events occurring at the early stages of fruit development that are crucial for the whole course of development. Some aspects of fruit quality (colour, shape, size, default) can be determined by looking at the outside of the fruit. Some other quality aspects (dry matter content, starch content, sweetness, etc..) must be invasively measured with specific tools and methods. Fruit quality is closely related to biochemical composition, mostly, although not only, due to low molecular-weight metabolites and the final fruit quality results from coordinated physiological processes during development from anthesis to growth and ripe stages [AGOA21]. In the last decade, we made major advances in metabolomics with the acquisition of profiles for a large number of metabolites of interest. Current metabolomics combines a range of analytical techniques for greater coverage of the metabolome. All the data generated can be further processed using statistical models in order to predict biomarkers of fruit quality or to help the production chain, from the seed to the plate of consumers.
For the fruit quality case, six different phenotypes have been predicted, including three categorical and three numerical variables (Growth Stage, fast VS slow growing, climacteric or not climacteric, relative growth rate, total acidity and total soluble sugars). Those six phenotypes have been predicted using metabolomic data acquired on fruit samples while RGR have been predicted using a combination of transcriptomic and metabolomic data.
Cassava is frequently described as a ‘food security’ crop primarily due to the fact that it can grow under adverse conditions but is nonetheless responsible for feeding over half a billion people. Cassava plants are typically propagated by cuttings, whereby a stem cutting from a stock plant is placed in the ground. This cutting quickly forms adventitious fibrous roots from the base of the cutting, which provide the plant with water and mineral salts or provides a support function. Some of these fibrous roots will undergo extensive swelling, thereby developing into storage roots. Cassava storage roots are analogous in function, such as the accumulation of food reserves, to the tuberous roots of sweet potato (Ipomea batatas) or the tubers of potato (Solanum tuberosum). In the case of cassava, the transition of a fibrous root to a storage root requires changes in both the function and morphology of the root to allow for the storage of large amounts of starch. The development of storage roots is believed to be influenced by factors such as the photoperiod and phytohormones. However, the transition of fibrous roots to storage roots is not fully understood [ChTa15]. Examining the RNA profile from different root types at different time points will provide us with a clearer picture of which genes are driving this transition.
Regarding the cassava case, information on root stage and age are available in relation to the development of the roots. Hence, we focused on the prediction of the different stages of cassava root development from fibrous roots to storage roots. Making such predictions from transcriptomics data enabled the discovery of gene biomarkers of the different stages, regardless of their age. We also performed predictions of the root ages to identify genes that could be markers of the time irrespective of their developmental stage.