RUVm (postprocessing method to reduce batch effects)

As with other types of microarray platforms, technical artifacts are a concern, including background fluorescence, dye-bias from the use of two color channels, bias caused by type I/II probe design, and batch effects. Several approaches and pipelines have been developed, either targeting a single issue or designed to address multiple biases through a combination of methods. We evaluate the effect of combining separate approaches to improve signal processing.

https://bioconductor.org/packages/release/bioc/vignettes/missMethyl/inst/doc/missMethyl.html#introduction

I found a new-ish method, RUVm, for removing unwanted variance between batches using negative controls. This works well after NOOB. Should I plan to port this from R into methylprep? https://cran.r-project.org/web/packages/ruv/ruv.pdf

Let’s hold off for now on RUVm for now. We may need to optimize our pre-processing pipeline, and need a strategy for empirically evaluating the utility of each step.

FoxoTech / methylprep

RUVm (postprocessing method to reduce batch effects) #33