Heritability estimation in high dimensional sparse linear mixed models

摘要: Motivated by applications in genetic fields, we propose to estimate the heritability high-dimensional sparse linear mixed models. The determines how variance is shared between different random components of a model. main novelty our approach consider that effects can be sparse, may contain null components, but do not know either their proportion or positions. estimator strongly inspired one proposed Pirinen, Donnelly and Spencer (2013), based on maximum likelihood approach. We also study theoretical properties estimator, namely establish root n-consistent when both number observations n N tend infinity under mild assumptions. prove satisfies central limit theorem which gives as byproduct confidence interval for heritability. Some Monte-Carlo experiments are conducted order show finite sample performances estimator.