Advancing genotype to phenotype Studies by considering TE Variability and Epivariabilty – STEVE

Genetic variation underpins inherited differences in traits, yet genome-wide association studies (GWAS) only explain part of this inheritance. One key explanation is that so far GWAS have largely neglected transposable element (TE) sequences, despite their prevalence and their often major effects on gene expression. Moreover, at least in plants and mammals, TE sequences are targeted by epigenetic mechanisms, notably DNA methylation, which potentially underlie an additional system of inheritance. However, their repetitive nature and their epigenetic properties pose unique challenges to calling their variants using short read sequencing data, and the inclusion of these variants in GWAS raises specific statistical difficulties, in particular due to the complexity of this type of variants. Here we propose through an interdisciplinary project to consider TE presence/absence and TE DNA methylation polymorphisms in genotype-to-phenotype studies, using as proof of principle the model plant Arabidopsis thaliana and novel statistical machine learning approaches. Our two main aims are to provide, (i) a comprehensive assessment and understanding of the epivariability of TE sequences; and (ii) a new methodological framework to conduct GWAS based on TE presence/absence variants as well as TE sequence epivariants. The findings of our project will have broad implications given the ubiquity of TE sequences in eukaryotic genomes and the substantial similarities between plants and mammals in the molecular machineries that underpin the epigenetic control of TEs. Our ambition is to provide a new foundation for dissecting out the genetic basis of complex traits, by tackling the complexities that TE (epi)variants bring to genetic studies, with direct applications for human health and agriculture.