| Summary: |
Portugal ranks worldwide among the most important wine producers in quantity, quality and diversity, and reached a trade record of 849M¤ in 2020. To face emerging competition
and challenges producers increasingly will depend on Precision Viticulture (PV). PV considers that both the amount and nature of spatio-temporal variations in vineyards are key
drivers for making highly-targeted decisions to increase productivity/quality, profitability, and minimize unintended environmental impacts. Current PV research is based on datadriven
captured by sophisticated sensors at airborne or proximal platforms, often benefiting from data-mining technologies to process big data[1]. But they do not provide what matters the most, ie, an in-situ diagnosis of vines physiology/metabolism, which allows predicting viticultural impacts driven by complex interaction of Genotype x Environment x
Management (GxExM). Also, they are not able to transpose the gathered wide information to new situations (eg, from one harvest to another, or from terroir to others)[3]. Major
challenges for future PV include, how to a) provide targeted and precision practices-based on plant physiology, and b) articulate them with state-of-the-art high-throughput (HTP)
multi-omics technologies (phenomics, transcriptomics, metabolomics)[5]. This allows using lab information to provide in real-time, a spatio-temporal-functional approach towards
improving PV-agricultural practices. This omics-integrative approach is easy to apply in controlled lab. environments but its translation to in-situ viticultural application remains a
challenge[7]. HTP applications in precision agriculture are emerging worldwide, and expensive HTP plant screening platforms are being established, enabling non-invasive measures
of omics data, but their application to vineyards is at its infancy[15,17]. For a breakthrough impact on PV and producers, it urges an innovation with low-cost platforms equipped
with photonic sensors for spati  |
Summary
Portugal ranks worldwide among the most important wine producers in quantity, quality and diversity, and reached a trade record of 849M¤ in 2020. To face emerging competition
and challenges producers increasingly will depend on Precision Viticulture (PV). PV considers that both the amount and nature of spatio-temporal variations in vineyards are key
drivers for making highly-targeted decisions to increase productivity/quality, profitability, and minimize unintended environmental impacts. Current PV research is based on datadriven
captured by sophisticated sensors at airborne or proximal platforms, often benefiting from data-mining technologies to process big data[1]. But they do not provide what matters the most, ie, an in-situ diagnosis of vines physiology/metabolism, which allows predicting viticultural impacts driven by complex interaction of Genotype x Environment x
Management (GxExM). Also, they are not able to transpose the gathered wide information to new situations (eg, from one harvest to another, or from terroir to others)[3]. Major
challenges for future PV include, how to a) provide targeted and precision practices-based on plant physiology, and b) articulate them with state-of-the-art high-throughput (HTP)
multi-omics technologies (phenomics, transcriptomics, metabolomics)[5]. This allows using lab information to provide in real-time, a spatio-temporal-functional approach towards
improving PV-agricultural practices. This omics-integrative approach is easy to apply in controlled lab. environments but its translation to in-situ viticultural application remains a
challenge[7]. HTP applications in precision agriculture are emerging worldwide, and expensive HTP plant screening platforms are being established, enabling non-invasive measures
of omics data, but their application to vineyards is at its infancy[15,17]. For a breakthrough impact on PV and producers, it urges an innovation with low-cost platforms equipped
with photonic sensors for spatio-temporal field-omics measurements. OmicBots is driven by 3 hypotheses: a) different conditions shift plants' omics data that are detectable by
spectral data; b) comparing lab vs. field multi-year-varieties-conditions gives insights on predominant GxExM factor; c) in-silico model provides a causal interpretation of omics
information, creating an information transfer platform between the lab and PV, for end-users. OmicBots aims to develop an automatic platform to integrate HTP low-cost roboticsassisted
smart-photonics sensors, with artificial intelligence and systems biology (in-silico), where grapevine metabolic pathways can be explored, to better comprehend how field
grapevine physiology/metabolism are driven by GxExM. This is crucial in viticulture, as wine industry is highly dependent on specific metabolites' signature (eg flavonoids), which
depend on GxExM[7]. Bioinformatics (eg CobraMatlab,22) combining both 'in-situ' phenotypes/metabolomics and lab. physiology/metabolomic/transcriptomic will be used to
implement an in-silico grapevine Genome-Scale Model (GSM) whereby the information between 'in-situ' and 'in-vitro' is bi-directionally transferable[23]. GSM is a mechanistic model
determining which genes, enzymes and metabolic pathways are activated, affecting plant physiology(fig1). Environmental impact on metabolome might be compared in lab. vs field
experiments (multi-year-varieties-conditions). We'll develop artificial intelligence algorithms for HTP putative annotation of metabolites (eg flavonoids), knowledge-based
downstream analysis, and validation of relevant grapevine-lik |