[Bug]: V2 Probe names need to be cleaned before EWAS

[rhaigh5]

Contact Details

[email protected]

What happened?

When using calcBrainComposition() function from testCellComposition in the EWAS pipeline, normalised data is not compatible due to new probe name changes in EPICv2. Fix: need to incorporate epicv2clean from wateRmelon into the normalisation script before saving data.

How can the bug be reproduced?

Run calcBrainComposition on normalised data produced from the QC pipeline.

Relevant log output

predPropBest<-calcBrainComposition(celltypeNormbeta)
Error in projectCellTypeWithError(betas, modelBrainCoef[[method]][[j]]) : 
  Only 0 probes in common between trained model and bulk data. Stopping.