Expand description
vooma: mean-variance modelling at the observational level for arrays.
Clean-room reimplementation of limma’s vooma() for a log-expression
matrix and design. Reference: Law CW (2013), “Precision weights for gene
expression analysis”, PhD Thesis, University of Melbourne
(hdl.handle.net/11343/38150). No limma (GPL) source was consulted; the
method follows the published documentation and is validated black-box
against the binary.
lmFit(y, design) gives per-gene residual sd (sigma) and row mean (Amean);
the trend is a lowess of sqrt(sigma) on Amean; each observation’s precision
weight is 1 / trend(fitted)^4 where fitted is its model-predicted mean.
Structs§
Functions§
- read_
design - Design TSV: col 1 = sample id, header = coefficient names, numeric model-matrix entries (one row per sample, in sample order).
- read_
expr - vooma
- write_
trend - write_
weights