ABI Innovation: A computational framework for integrating image informatics with transcriptomics for discovering spatiotemporally resolved regulatory gene networks in plants

University Of Nebraska-Lincoln, Lincoln NE

Investigators

Abstract

Biological processes at all scales from molecules to ecosystems are coordinated through the encoding, exchange, and interpretation of information. Many of the recent advances in the biological sciences collect very large sets of DNA (genotype) and protein sequence: making sense of how their properties lead to the appearance, responses and behaviors (phenotype) of organisms requires new computational tools. Linking the genotype to the phenotype requires untangling very complicated relationships. Collecting large amounts of data for large numbers of samples means developing new tools for data processing and data modeling. This project will develop computational tools to address both of these challenges, while investigating the effects of salt tolerance in rice. Large sets of expression data will be collected along with images of the developing plants from several imaging technologies. The images will be used to produce 3D reconstructions of the plant architecture that can be stored as digital phenotypes. It is expected that these digital phenotypes will allow plant breeders to identify traits that were not captured by standard observation methods, and find new genotype-phenotype associations. The educational activities are timely, providing training opportunities for plant biology and computer science students. As the volumes of image-derived digital data surge, image informatics is emerging as a cutting-edge discovery tool. Outreach and training activities include a summer workshop for plant phenomics and establishing a data visualization community at University of Nebraska. Project resources will be released on CyVerse to ensure wide availability. High-throughput sequencing technology- driven transcriptome analyses are a commonly utilized approach for gaining molecular insights on an organism's response to environmental changes. These transcriptome-level responses and the underlying regulatory gene networks that drive phenotypic responses to changing environment are highly dynamic. With the advent of high-throughput image-based plant phenotyping, it is now possible to capture the dynamics of growth and other digital-features responding to environmental or genetic perturbation with increased temporal and spatial resolution. This project will use temporal imaging to inform the collection of time series transcriptome data with 3D-spatial sensitivity for discovering dynamic regulatory co-expression networks. Outcomes of the project will be to develop innovative algorithms to integrate heterogeneous datasets and implement it in a prototype computational framework. The platform will enable interactive, multidimensional visualization of dynamic regulatory networks with spatial and temporal resolution. The project resources can be accessed at: http://cropstressgenomics.org/phenomics.php