Biology subjects

Ramstein, G.

Publications and source records attributed to Ramstein, G..

4 recordsLinked to original sources

bioRxiv2025

Genomic prediction of agronomic traits in perennial ryegrass (Lolium perenne L.) and genotype x environment interactions at the limit of the species distribution

BackgroundIn breeding the aim is to identify and accumulate beneficial variants. However, detection of these variants may be challenging in the presence of extensive genotype x environment interactions (GxE), as variant effects will be conditional on environment. MethodsThe study assesses the performance of 264 diploid perennial ryegrass accessions in a multi-environment field trial. We investigate the extent of GxE, for yield (total dry matter) and persistence traits, i.e. winter kill and spring cover, under environmental conditions experienced in Nordic and Baltic regions at the limit of the species distribution. Two different approaches to modelling GxE were tested: reaction norm and envirotyping, and the models were validated under three different breeding scenarios. ResultsOur analysis documented the presence of significant GxE interaction for all traits investigated in the study. Validation showed improvements in prediction accuracy when accounting for GxE: up to 4% for yield when predicting in unobserved environments, and up to 22% and 9% for spring cover and winter kill, respectively when predicting unobserved germplasm. Genome-wide-association-studies (GWAS) were utilized to detect genetic variants with marginal effects (environment-independent effect) and conditional effects (environment-dependent effects). Results showed the presence of large-effect genetic variants with marginal effects, in addition to few QTL whose effects were adaptive under specific environmental conditions while neutral or deleterious under different environmental conditions. ConclusionThis study demonstrates the usefulness and limitations of genomic prediction models for predicting GxE in highly diverse samples, and describes the extent of GxE interactions at the limit of species distribution for perennial ryegrass. Finally, our study points toward adaptive variation, which may enhance persistence of perennial ryegrass populations in Nordic and Baltic growing conditions.

genetics↗

bioRxiv2021

Predicting phenotypes from genetic, environment, management, and historical data using CNNs

Predicting phenotypes from genetic (G), environmental (E), and management (M) conditions is a long-standing challenge with implications to agriculture, medicine, and conservation. Most methods reduce the factors in a dataset (feature engineering) in a subjective and potentially oversimplified manner. Convolutional Neural Networks (CNN) can overcome this by allowing the data itself to determine which factors are most important. CNN models were developed for predicting agronomic yield from a combination of replicated trials and historical yield survey data. The results were more accurate than standard methods when tested on heldout G, E, and M data (r=0.5 vs r=0.4), and performed slightly worse than standard methods when only G was held out (r=0.74 vs r=0.78). Pre-training on historical data increased accuracy by 1-36% compared to trial data alone. Saliency map analysis indicated the CNN has "learned" to prioritize many factors of known agricultural importance.

plant biology↗

bioRxiv2020

FaDA: A Shiny web application to accelerate common lab data analyses

BackgroundWeb-based data analysis and visualization tools are mostly designed for specific purposes, such as data from whole transcriptome RNA sequencing or single-cell RNA sequencing. However, limited efforts have been made to develop tools designed for data of common laboratory data for non-computational scientists. The importance of such web-based tool is stressed by the current increased samples capacity of conventional laboratory tools such as quantitative PCR, flow cytometry or ELISA. ResultsWe provide a web-based application FaDA, developed with the R Shiny package providing users to perform statistical group comparisons, including parametric and non-parametric tests, with multiple testing corrections suitable for most of the standard wet-lab analyses. FaDA provides data visualization such as heatmap, principal component analysis (PCA) and receiver operating curve (ROC). Calculations are performed through the R language. ConclusionsFaDA application provides a free and intuitive interface allowing biologists without bioinformatic skills to easily and quickly perform common lab data analyses. The application is freely accessible at https://shiny-bird.univ-nantes.fr/app/Fada AbbreviationsAUC: Area Under the Curve; FaDA: Fast Data Analysis; GEO: Gene Expression Omnibus; ELISA: enzyme-linked immunosorbent assay; PCA: Principal Component Analysis; qPCR: quantitative PCR; ROC: Receiver Operating Curve.

bioinformatics↗

bioRxiv2018

Evolutionarily informed deep learning methods: Predicting transcript abundance from DNA sequence

Deep learning methodologies have revolutionized prediction in many fields, and show potential to do the same in molecular biology and genetics. However, applying these methods in their current forms ignores evolutionary dependencies within biological systems and can result in false positives and spurious conclusions. We developed two novel approaches that account for evolutionary relatedness in machine learning models: 1) gene-family guided splitting, and 2) ortholog contrasts. The first approach accounts for evolution by constraining the models training and testing sets to include different gene families. The second, uses evolutionarily informed comparisons between orthologous genes to both control for and leverage evolutionary divergence during the training process. The two approaches were explored and validated within the context of mRNA expression level prediction, and have prediction auROC values ranging from 0.72 to 0.94. Model weight inspections showed biologically interpretable patterns, resulting in the novel hypothesis that the 3 UTR is more important for fine tuning mRNA abundance levels while the 5 UTR is more important for large scale changes.

molecular biology↗