In the early, online edition of the Proceedings of the National Academy of Sciences, a Chinese Academy of Sciences-led team takes a look at DNA methylation patterns in hybrid Arabidopsis plants. Using bisulfite sequencing and small RNA sequencing, the researchers characterized the DNA methylomes and small RNA repertoires in two different Arabidopsis ecotype plants and hybrid progeny formed by crossing them. With the help of weighted methylation analyses that considered sequence variations between the parental plants, they identified parts of the genome showing methylation interactions in the hybrid Arabidopsis progeny, including some 2,500 with similar methylation levels as parental plants and about 1,000 interactions at differentially methylated regions.
Researchers from Microsoft, the UK, and Uganda assess computational models for estimating heritability that take into account random effects such as environmental variation. The team's strategy was to focus on a more general version of the linear mixed model as a means of trying to curb inflations in heritability estimates. The group used the general LMM with either random genomic variant effects or random environmental spatial location effects to evaluated simulated data or data for nearly 4,800 individuals from Uganda who were evaluated for dozens of phenotypes, including blood pressure and glycemic control. Results from the analyses revealed lower heritability estimates with the more general model, which the study's authors say might partly explain the missing heritability problem described in the past.
Finally, a team from Switzerland, the Netherlands, and Germany introduces so-called invariant causal prediction, or ICP, software for inferring causal features for a phenotype of interest by quantifying confidence probabilities in causal structure information with statistical inference. The researchers applied ICP to messenger RNA expression levels and other data from large-scale gene deletion experiments in the budding yeast, Saccharomyces cerevisiae, for example, uncovering what appeared to be relatively reliable causal relationships. They found that the approach also uncovered promising results in experiments focused on finding biochemical interactions in yeast under different experimental conditions or interventions.