Weighted matrix manipulation, finding components of variation with weighted or sparse data
Data type and tools for working with matrices having precision weights and missing data. This package provides a common representation and tools that can be used with many types of high-throughput data. The meaning of the weights is compatible with usage in the base R function "lm" and the package "limma". Calibrate weights by scaling weights row-wise to account for known predictors of precision. Find PCA-like components of variation even with many missing values, rotated so that individual components may be meaningfully interpreted. DelayedArray matrices and BiocParallel are supported.