Search bioRxiv⌕ Search
Home/Archive/bioRxiv

bioRxiv · 10.1101/071662

Deep learning based root-soil segmentation from X-ray tomography

Clement DOUARRE·Richard SCHIELEIN·Carole FRINDEL·Stefan GERTH·David ROUSSEAU
View original source ↗Export metadata ↓

Abstract

One of the most challenging computer vision problem in plant sciences is the segmentation of root and soil from X-ray tomography. So far, this has been addressed from classical image analysis methods. In this paper, we address this root/soil segmentation problem from X-ray tomography using a new deep learning classification technique. The robustness of this technique, tested for the first time on this plant science problem, is established with root/soil presenting a very low contrast in X-ray tomography. We also demonstrate the possibility to segment efficiently root from soil while learning on purely synthetic soil and root.

Explore related subjects

Plant BiologyPreprint

Keep this discovery

BibTeXRIS

Clement DOUARRE, Richard SCHIELEIN, Carole FRINDEL, Stefan GERTH, David ROUSSEAU. 2016-08-25. Deep learning based root-soil segmentation from X-ray tomography. https://doi.org/10.1101/071662

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2016

The PhytoClust Tool for Metabolic Gene Clusters Discovery in Plant Genomes

The existence of Metabolic Gene Clusters (MGCs) in plant genomes has recently raised increased interest. Thus far, MGCs were commonly identified for pathways of specialized metabolism, mostly those associated with terpene type products. For efficient identification of novel MGCs computational approaches are essential. Here we present PhytoClust; a tool for the detection of candidate MGCs in plant genomes. The algorithm employs a collection of enzyme families related to plant specialized metabolism, translated into hidden Markov models, to mine given genome sequences for physically co-localized metabolic enzymes. Our tool accurately identifies previously characterized plant MBCs. An exhaustive search of 31 plant genomes detected 1232 and 5531 putative gene cluster types and candidates, respectively. Clustering analysis of putative MGCs types by species reflected plant taxonomy. Furthermore, enrichment analysis revealed taxa- and species-specific enrichment of certain enzyme families in MGCs. When operating through our web-interface, PhytoClust users can mine a genome either based on a list of known cluster types or by defining new cluster rules. Moreover, for selected plant species, the output can be complemented by co-expression analysis. Altogether, we envisage PhytoClust to enhance novel MGCs discovery which will in turn impact the exploration of plant metabolism.

Plant Biology↗
bioRxiv2016

Morphological plant modeling: Unleashing geometric and topologic potential within the plant sciences

Plant morphology is inherently mathematical in that morphology describes plant form and architecture with geometrical and topological descriptors. The geometries and topologies of leaves, flowers, roots, shoots and their spatial arrangements have fascinated plant biologists and mathematicians alike. Beyond providing aesthetic inspiration, quantifying plant morphology has become pressing in an era of climate change and a growing human population. Modifying plant morphology, through molecular biology and breeding, aided by a mathematical perspective, is critical to improving agriculture, and the monitoring of ecosystems with fewer natural resources. In this white paper, we begin with an overview of the mathematical models applied to quantify patterning in plants. We then explore fundamental challenges that remain unanswered concerning plant morphology, from the barriers preventing the prediction of phenotype from genotype to modeling the movement of leafs in air streams. We end with a discussion concerning the incorporation of plant morphology into educational programs. This strategy focuses on synthesizing biological and mathematical approaches and ways to facilitate research advances through outreach, cross-disciplinary training, and open science. This white paper arose from bringing mathematicians and biologists together at the National Institute for Mathematical and Biological Synthesis (NIMBioS) workshop titled \"Morphological Plant Modeling: Unleashing Geometric and Topological Potential within the Plant Sciences\" held at the University of Tennessee, Knoxville in September, 2015. Never has the need to quantify plant morphology been more imperative. Unleashing the potential of geometric and topological approaches in the plant sciences promises to transform our understanding of both plants and mathematics.

Plant Biology↗
bioRxiv2016

Sequencing of small RNAs of the fern Pleopeltis minima (Polypodiaceae) offers insight into the evolution of the microRNA repertoire in land plants

MicroRNAs (miRNAs) are short, single stranded RNA molecules that regulate the stability and translation of messenger RNAs in diverse eukaryotic groups. Several miRNA genes are of ancient origin and have been maintained in the genomes of animal and plant taxa for hundreds of millions of years, and functional studies indicate that ancient miRNAs play key roles in development and physiology. In the last decade, genome and small RNA (sRNA) sequencing of several plant species have helped unveil the evolutionary history of land plant miRNAs. Land plants are divided into bryophytes (liverworts, mosses), lycopods (clubmosses and spikemosses), monilophytes (ferns and horsetails), gymnosperms (cycads, conifers and allies) and angiosperms (flowering plants). Among these, the fern group occupies a key phylogenetic position, since it represents the closest extant cousin taxon of seed plants, i.e. gymno- and angiosperms. However, in spite of their evolutionary, economic and ecological importance, no fern genome has been sequenced yet and few genomic resources are available for this group. Here, we sequenced the small RNA fraction of an epiphytic South American fern, Pleopeltis minima (Polypodiaceae), and compared it to plant miRNA databases, allowing for the identification of miRNA families that are shared by all land plants, shared by all vascular plants (tracheophytes) or shared by euphyllophytes (ferns and seed plants) only. Using the recently described transcriptome of another fern, Lygodium japonicum, we also estimated the degree of conservation of fern miRNA targets in relation to other plant groups. Our results pinpoint the origin of several miRNA families in the land plant evolutionary tree with more precision and are a resource for future genomic and functional studies of fern miRNAs.

Plant Biology↗