Limits...
Crop improvement using life cycle datasets acquired under field conditions.

Mochida K, Saisho D, Hirayama T - Front Plant Sci (2015)

Bottom Line: Recent advances in omics analytical techniques and information technology have enabled us to integrate data from a spectrum of physiological metrics of field crops.Here, we describe recent advances in technologies that are key components for data driven crop design, such as population genomics, chronological omics analyses, and computer-aided molecular network prediction.These elements would help us to genetically engineer "designed crops" to prevent yield shortfalls because of environmental fluctuations due to future climate change.

View Article: PubMed Central - PubMed

Affiliation: Cellulose Production Research Team, Biomass Engineering Research Division, RIKEN Center for Sustainable Resource Science , Yokohama, Japan ; Gene Discovery Research Group, RIKEN Center for Sustainable Resource Science , Yokohama, Japan ; Kihara Institute for Biological Research, Yokohama City University , Yokohama, Japan.

ABSTRACT
Crops are exposed to various environmental stresses in the field throughout their life cycle. Modern plant science has provided remarkable insights into the molecular networks of plant stress responses in laboratory conditions, but the responses of different crops to environmental stresses in the field need to be elucidated. Recent advances in omics analytical techniques and information technology have enabled us to integrate data from a spectrum of physiological metrics of field crops. The interdisciplinary efforts of plant science and data science enable us to explore factors that affect crop productivity and identify stress tolerance-related genes and alleles. Here, we describe recent advances in technologies that are key components for data driven crop design, such as population genomics, chronological omics analyses, and computer-aided molecular network prediction. Integration of the outcomes from these technologies will accelerate our understanding of crop phenology under practical field situations and identify key characteristics to represent crop stress status. These elements would help us to genetically engineer "designed crops" to prevent yield shortfalls because of environmental fluctuations due to future climate change.

No MeSH data available.


Related in: MedlinePlus

A conceptual framework for phenology data driven crop design. Phenology datasets include plant phenotypes, omics snapshots with environmental data. A genome-wide polymorphism dataset is useful to find genetic association between accessions with traits as well as with physiological state. For model building, the collected data are applied to various types of computer-aided methods. Digitized datasets can also be assessed against data on well-characterized gene functions in model plants. Machine learning-based data clustering and network prediction may help us to identify candidate genes and alleles for crop improvement.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4585263&req=5

Figure 2: A conceptual framework for phenology data driven crop design. Phenology datasets include plant phenotypes, omics snapshots with environmental data. A genome-wide polymorphism dataset is useful to find genetic association between accessions with traits as well as with physiological state. For model building, the collected data are applied to various types of computer-aided methods. Digitized datasets can also be assessed against data on well-characterized gene functions in model plants. Machine learning-based data clustering and network prediction may help us to identify candidate genes and alleles for crop improvement.

Mentions: Machine learning is a field of computer science for the design of computational algorithms that automatically improve with experience. In the last two decades, this research field has dramatically advanced with the emergence of artificial intelligence and data science, and has been applied in various fields in science, technology and commerce (Jordan and Mitchell, 2015). Machine learning is also used in applications for the analysis of genome-scale datasets and other large-scale omics datasets in life science (Libbrecht and Noble, 2015). Learning methods in machine learning are usually classified into two primary categories of supervised and unsupervised learning. The supervised machine learning aims to produce an algorithm to predict output on unknown input via a training process using a dataset of known pairs of input and output. The unsupervised machine learning methods are used to extract structures and identify their features in a given dataset without examples for training. Computational modeling using machine learning has been performed recently with the aim of predicting gene networks based on large-scale transcriptome datasets in plants. For example, in Arabidopsis, supervised machine learning was used to build a network model of responses to stress conditions, to explore genes related to stress responses, and to predict molecular interactions (Ma et al., 2014; Nourani et al., 2015). Machine learning provides a data-driven approach to extract latent rules or patterns from a comprehensively collected dataset without any biased view on the biological phenomena of interest (Figure 2).


Crop improvement using life cycle datasets acquired under field conditions.

Mochida K, Saisho D, Hirayama T - Front Plant Sci (2015)

A conceptual framework for phenology data driven crop design. Phenology datasets include plant phenotypes, omics snapshots with environmental data. A genome-wide polymorphism dataset is useful to find genetic association between accessions with traits as well as with physiological state. For model building, the collected data are applied to various types of computer-aided methods. Digitized datasets can also be assessed against data on well-characterized gene functions in model plants. Machine learning-based data clustering and network prediction may help us to identify candidate genes and alleles for crop improvement.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4585263&req=5

Figure 2: A conceptual framework for phenology data driven crop design. Phenology datasets include plant phenotypes, omics snapshots with environmental data. A genome-wide polymorphism dataset is useful to find genetic association between accessions with traits as well as with physiological state. For model building, the collected data are applied to various types of computer-aided methods. Digitized datasets can also be assessed against data on well-characterized gene functions in model plants. Machine learning-based data clustering and network prediction may help us to identify candidate genes and alleles for crop improvement.
Mentions: Machine learning is a field of computer science for the design of computational algorithms that automatically improve with experience. In the last two decades, this research field has dramatically advanced with the emergence of artificial intelligence and data science, and has been applied in various fields in science, technology and commerce (Jordan and Mitchell, 2015). Machine learning is also used in applications for the analysis of genome-scale datasets and other large-scale omics datasets in life science (Libbrecht and Noble, 2015). Learning methods in machine learning are usually classified into two primary categories of supervised and unsupervised learning. The supervised machine learning aims to produce an algorithm to predict output on unknown input via a training process using a dataset of known pairs of input and output. The unsupervised machine learning methods are used to extract structures and identify their features in a given dataset without examples for training. Computational modeling using machine learning has been performed recently with the aim of predicting gene networks based on large-scale transcriptome datasets in plants. For example, in Arabidopsis, supervised machine learning was used to build a network model of responses to stress conditions, to explore genes related to stress responses, and to predict molecular interactions (Ma et al., 2014; Nourani et al., 2015). Machine learning provides a data-driven approach to extract latent rules or patterns from a comprehensively collected dataset without any biased view on the biological phenomena of interest (Figure 2).

Bottom Line: Recent advances in omics analytical techniques and information technology have enabled us to integrate data from a spectrum of physiological metrics of field crops.Here, we describe recent advances in technologies that are key components for data driven crop design, such as population genomics, chronological omics analyses, and computer-aided molecular network prediction.These elements would help us to genetically engineer "designed crops" to prevent yield shortfalls because of environmental fluctuations due to future climate change.

View Article: PubMed Central - PubMed

Affiliation: Cellulose Production Research Team, Biomass Engineering Research Division, RIKEN Center for Sustainable Resource Science , Yokohama, Japan ; Gene Discovery Research Group, RIKEN Center for Sustainable Resource Science , Yokohama, Japan ; Kihara Institute for Biological Research, Yokohama City University , Yokohama, Japan.

ABSTRACT
Crops are exposed to various environmental stresses in the field throughout their life cycle. Modern plant science has provided remarkable insights into the molecular networks of plant stress responses in laboratory conditions, but the responses of different crops to environmental stresses in the field need to be elucidated. Recent advances in omics analytical techniques and information technology have enabled us to integrate data from a spectrum of physiological metrics of field crops. The interdisciplinary efforts of plant science and data science enable us to explore factors that affect crop productivity and identify stress tolerance-related genes and alleles. Here, we describe recent advances in technologies that are key components for data driven crop design, such as population genomics, chronological omics analyses, and computer-aided molecular network prediction. Integration of the outcomes from these technologies will accelerate our understanding of crop phenology under practical field situations and identify key characteristics to represent crop stress status. These elements would help us to genetically engineer "designed crops" to prevent yield shortfalls because of environmental fluctuations due to future climate change.

No MeSH data available.


Related in: MedlinePlus