In this episode Dominik gives us insights into CropML, a BMBF funded project. The project evaluates new machine learning techniques for more accurate plant breeding by integrating heterogeneous external factors. Different phenotype prediction models, including basic genomic selection methods to more advanced deep learning-based techniques have been compared. Learn why advanced models are the future and where the challenges are.
New paper in Scientific Data: “HeliantHOME, a public and centralized database of phenotypic sunflower data”. We have created HeliantHOME (http://www.helianthome.org), a curated, public, and interactive database of phenotypes including developmental, structural and environmental ones, obtained from a large collection of both wild and cultivated sunflower individuals.
New paper: “A comparison of classical and machine learning-based phenotype prediction methods on simulated data and three plant species”. In this work, our aim is to systematically compare 12 different phenotype prediction models, including basic genomic selection methods to more advanced deep learning-based techniques. More importantly, we assess the performance of these models on simulated phenotype data as well as on real-world data from Arabidopsis thaliana and two breeding datasets from soy and corn. The synthetic phenotypic data allow us to analyze all prediction models and especially the selected markers under controlled and predefined settings. We show that Bayes B and linear regression models with sparsity constraints perform best under different simulation settings with respect to explained variance. Further, we can confirm results from other studies that there is no superiority of more complex neural network-based architectures for phenotype prediction compared to well-established methods. However, on real-world data, for which several prediction models yield comparable results with slight advantages for Elastic Net, this picture is less clear, suggesting that there is a lot of room for future research.
New paper about “Deep learning-based early weed segmentation using motion blurred UAV images of sorghum fields”. In this work, we developed and analysed different deep learning-based architectures to accurately segment crop and weed species in UAV images of agricultural fields under challenging capturing conditions. For this propose, we created an expert-curated fully-annotated weed segmentation UAV dataset in sorghum fields. We show that our trained models have high precision in detecting general plant shapes with minor weaknesses at borders of the plants. More importantly, our method is capable in segmenting intra-row and partly occluded weeds on an individual plant basis. All code and data are publicly available on GitHub and Mendeley Data.
New paper about “Systematic analysis of the underlying genomic architecture for transcriptional–translational coupling in prokaryotes”. In this work, we systematically analyzed gene cassettes from more than 1800 bacterial for the abundance of transcriptional and translational associated genes clustered in distinct gene cassettes. We identified three highly frequent cassettes containing transcriptional and translational genes. Interestingly, each of the three cassettes harbors a gene (nusG, rpsD and nusA) encoding a protein which links transcription and translation in bacteria. Furthermore, our analyses suggest an enrichment of these gene cassettes in pathogenic bacterial phyla.
Maura presents our new paper on “Efficient permutation-based genome-wide association studies for normal and skewed phenotypic distributions” at the 21st European Conference on Computational Biology (ECCB), the largest bioinformatics conference in Europe.