Betavalue

Betavalue is a term used in several disciplines to denote a parameter or derived quantity associated with probability models, statistics, or data analysis. Because meanings vary by field, the term often refers to a regression coefficient, a methylation proportion in genomics, or a probability parameter in beta distributions.

In regression analysis, the beta value (beta coefficient) measures the effect size of a predictor on the response variable. An unstandardized beta reflects change in the dependent variable per unit change in the predictor and depends on the measurement scales. A standardized beta expresses effects in units of standard deviations, enabling comparison across predictors.

In epigenetics and DNA methylation studies, the beta value denotes the proportion of methylation at a CpG site. It is commonly defined as beta = M / (M + U + alpha), where M and U are methylated and unmethylated signal intensities and alpha is a small constant to stabilize the ratio. Values range from 0 to 1. Analysts often transform to M-values for statistical testing, because they have more favorable statistical properties.

In Bayesian statistics, the beta distribution models probabilities on the interval [0,1], with two shape parameters alpha and beta. In this context, betavalue may refer to a drawn probability or to a model parameter that follows a Beta distribution, such as a Bayesian estimate of a proportion or a Beta-Binomial conjugate prior.