Boosting heritability

"Boosting heritability" is a software tool to enhance the precision and reliability of heritability estimates in genetic studies. Heritability, a cornerstone concept in genetics, measures the proportion of observed variability in a trait attributed to genetic differences among individuals. Traditional approaches to estimating heritability often rely on random-effect models, primarily for computational convenience. However, these methods can sometimes lack the precision and adaptability required for complex genetic data, mainly when dealing with high-dimensional datasets.

In contrast, "boosting heritability" employs a fixed-effect model approach, which has historically been less explored due to its computational complexity. This strategy integrates the strengths of various recent methodologies, employing a high-dimensional linear model to estimate heritability. One of the critical innovations of "boosting heritability" is its use of a multiple-sample splitting strategy. This approach divides the dataset into several subsamples, performs estimation on each, and then combines these estimates to achieve a final heritability estimate. This technique enhances the stability and accuracy of the heritability estimate, making it particularly effective for analyzing complex and high-dimensional genetic data.