Deep phenotyping: deep learning for temporal phenotype/genotype classification.

BACKGROUND: High resolution and high throughput genotype to phenotype studies in plants are underway to accelerate breeding of climate ready crops. In the recent years, deep learning techniques and in particular Convolutional Neural Networks (CNNs), Recurrent Neural Networks and Long-Short Term Memories (LSTMs), have shown great success in visual data recognition, classification, and sequence learning tasks. More recently, CNNs have been used for plant classification and phenotyping, using individual static images of the plants. On the other hand, dynamic behavior of the plants as well as their growth has been an important phenotype for plant biologists, and this motivated us to study the potential of LSTMs in encoding these temporal information for the accession classification task, which is useful in automation of plant production and care. METHODS: In this paper, we propose a CNN-LSTM framework for plant classification of various genotypes. Here, we exploit the power of deep CNNs for automatic joint feature and classifier learning, compared to using hand-crafted features. In addition, we leverage the potential of LSTMs to study the growth of the plants and their dynamic behaviors as important discriminative phenotypes for accession classification. Moreover, we collected a dataset of time-series image sequences of four accessions of Arabidopsis, captured in similar imaging conditions, which could be used as a standard benchmark by researchers in the field. We made this dataset publicly available. CONCLUSION: The results provide evidence of the benefits of our accession classification approach over using traditional hand-crafted image analysis features and other accession classification frameworks. We also demonstrate that utilizing temporal information using LSTMs can further improve the performance of the system. The proposed framework can be used in other applications such as in plant classification given the environment conditions or in distinguishing diseased plants from healthy ones.

Variation in Leaf Respiration Rates at Night Correlates with Carbohydrate and Amino Acid Supply.

Plant respiration can theoretically be fueled by and dependent upon an array of central metabolism components; however, which ones are responsible for the quantitative variation found in respiratory rates is unknown. Here, large-scale screens revealed 2-fold variation in nighttime leaf respiration rate (RN) among mature leaves from an Arabidopsis (Arabidopsis thaliana) natural accession collection grown under common favorable conditions. RN variation was mostly maintained in the absence of genetic variation, which emphasized the low heritability of RN and its plasticity toward relatively small environmental differences within the sampling regime. To pursue metabolic explanations for leaf RN variation, parallel metabolite level profiling and assays of total protein and starch were performed. Within an accession, RN correlated strongly with stored carbon substrates, including starch and dicarboxylic acids, as well as sucrose, major amino acids, shikimate, and salicylic acid. Among different accessions, metabolite-RN correlations were maintained with protein, sucrose, and major amino acids but not stored carbon substrates. A complementary screen of the effect of exogenous metabolites and effectors on leaf RN revealed that (1) RN is stimulated by the uncoupler FCCP and high levels of substrates, demonstrating that both adenylate turnover and substrate supply can limit leaf RN, and (2) inorganic nitrogen did not stimulate RN, consistent with limited nighttime nitrogen assimilation. Simultaneous measurements of RN and protein synthesis revealed that these processes were largely uncorrelated in mature leaves. These results indicate that differences in preceding daytime metabolic activities are the major source of variation in mature leaf RN under favorable controlled conditions.

TraitCapture: genomic and environment modelling of plant phenomic data.

Agriculture requires a second green revolution to provide increased food, fodder, fiber, fuel and soil fertility for a growing population while being more resilient to extreme weather on finite land, water, and nutrient resources. Advances in phenomics, genomics and environmental control/sensing can now be used to directly select yield and resilience traits from large collections of germplasm if software can integrate among the technologies. Traits could be Captured throughout development and across environments from multi-dimensional phenotypes, by applying Genome Wide Association Studies (GWAS) to identify causal genes and background variation and functional structural plant models (FSPMs) to predict plant growth and reproduction in target environments. TraitCapture should be applicable to both controlled and field environments and would allow breeders to simulate regional variety trials to pre-select for increased productivity under challenging environments.