Biology subjects

Spangenberg, G.

Publications and source records attributed to Spangenberg, G..

4 recordsLinked to original sources

bioRxiv2026

DeepBioGS: a hybrid framework for integrating crop growth modelling with genomic prediction through neural networks

Ensuring global food security under rapid climate change demands accelerated genetic gain and breeding strategies that address complex Genotype-by-Environment (GxE) interactions. Traditional genomic selection models often fail to account for novel or extreme climates.Furthermore, integrating mechanistic crop growth models (CGMs) using traditional Bayesian frameworks to solve this issue presents severe computational bottlenecks. Here, we introduce DeepBioGS, a novel hybrid framework that integrates genomic selection with biophysical growth modelling via a fully differentiable deep learning architecture. DeepBioGS utilises a parameter-prediction multi-layer perceptron to map high-dimensional genomic markers to latent, highly heritable physiological traits (Genotype-Specific Parameters; GSP). These parameters mechanistically predict crop phenology across diverse environments. Using two multi-environment wheat datasets comprising over 6,000 genotypes, DeepBioGS extracted latent traits with near-perfect SNP-based heritability values (0.95-1.00). Crucially, the framework demonstrated superior or comparable predictive accuracy (up to r2 = 0.77) against standard genomic best linear unbiased prediction (GBLUP) and traditional Bayesian CGM-WGP models. Its architecture drastically improved computational scalability by enabling standard backpropagation, effectively bypassing the stochastic sampling limitations of approximate Bayesian methods. Most importantly for climate adaptation, DeepBioGS allowed accurate forecasting of genotype performance in entirely unobserved environmental conditions. By merging the representational power of deep learning with the structural constraints of biophysics, DeepBioGS provides a highly scalable, interpretable tool to navigate GxE interactions, enabling the assessment of cultivars under future climate scenarios, thus optimising crop breeding for a changing global environment.

plant biology↗

bioRxiv2021

Automated Hyperspectral Vegetation Index Derivation Using A Hyperparameter Optimization Framework for High-Throughput Plant Phenotyping

O_LIHyperspectral vegetation indices (VIs) are widely deployed in agriculture remote sensing and plant phenotyping to estimate plant biophysical and biochemical traits. However, existing VIs consist mainly of simple 2-band indices which limits the net performance and often do not generalize well for other traits than they were originally designed for. C_LIO_LIWe present an automated hyperspectral vegetation index (AutoVI) system for the rapid generation of novel 2- to 6-band trait-specific indices in a streamlined process covering model selection, optimization and evaluation driven by the tree parzen estimator algorithm. Its performance was tested in generating novel indices to estimate chlorophyll and sugar contents in wheat. C_LIO_LIResults show that AutoVI can rapidly generate complex novel VIs ([≥]4-band index) which correlated strongly (R2 > 0.8) with measured chlorophyll and sugar contents in wheat. AutoVI-derived indices were used as features in simple and stepwise multiple linear regression for chlorophyll and sugar content estimation, and outperformed results achieved with existing 47 VIs and those provided by partial least squares regression. C_LIO_LIThe AutoVI system can deliver novel trait-specific VIs readily adoptable in high-throughput plant phenotyping platforms and should appeal to plant scientists and breeders. A graphical user interface of AutoVI is herein provided. C_LI

plant biology↗

bioRxiv2020

A New and Improved Genome Sequence of Cannabis sativa

Cannabis is a diploid species (2n = 20), the estimated haploid genome sizes of the female and male plants using flow cytometry are 818 and 843 Mb respectively. Although the genome of Cannabis has been sequenced (from hemp, wild and high-THC strains), all assemblies have significant gaps. In addition, there are inconsistencies in the chromosome numbering which limits their use. A new comprehensive draft genome sequence assembly (~900 Mb) has been generated from the medicinal cannabis strain Cannbio-2, that produces a balanced ratio of cannabidiol and delta-9-tetrahydrocannabinol using long-read sequencing. The assembly was subsequently analysed for completeness by ordering the contigs into chromosome-scale pseudomolecules using a reference genome assembly approach, annotated and compared to other existing reference genome assemblies. The Cannbio-2 genome sequence assembly was found to be the most complete genome sequence available based on nucleotides assembled and BUSCO evaluation in Cannabis sativa with a comprehensive genome annotation. The new draft genome sequence is an advancement in Cannabis genomics permitting pan-genome analysis, genomic selection as well as genome editing.

genomics↗

bioRxiv2020

Automated Machine Learning for High-Throughput Image-Based Plant Phenotyping

Automated machine learning (AutoML) has been heralded as the next wave in artificial intelligence with its promise to deliver high performance end-to-end machine learning pipelines with minimal effort from the user. AutoML with neural architecture search which searches for the best neural network architectures in deep learning has delivered state-of-the-art performance in computer vision tasks such as image classification and object detection. Using wheat lodging assessment with UAV imagery as an example, we compared the performance of an open-source AutoML framework, AutoKeras in image classification and regression tasks to transfer learning using modern convolutional neural network (CNN) architectures pretrained on the ImageNet dataset. For image classification, transfer learning with Xception and DenseNet-201 achieved best classification accuracy of 93.2%, whereas Autokeras had 92.4% accuracy. For image regression, transfer learning with DenseNet-201 had the best performance (R2=0.8303, RMSE=9.55, MAE=7.03, MAPE=12.54%), followed closely by AutoKeras (R2=0.8273, RMSE=10.65, MAE=8.24, MAPE=13.87%). Interestingly, in both tasks, AutoKeras generated compact CNN models with up to 40-fold faster inference times compared to the pretrained CNNs. The merits and drawbacks of AutoML compared to transfer learning for image-based plant phenotyping are discussed.

plant biology↗