Hi,
I am working on RNAseq data where my colleague has used two ERCC mixes , Mix 1 for tests and Mix 2 for WT. https://www.thermofisher.com/order/catalog/product/4456740 So, I have 92 ERCC transcripts with different expected fold change. I am not sure how to provide the information for two mixes when I normalize the data using loess.normalization. I know I should provide a subset of rows (ERCC transcript) and perform normalization to the whole matrix in R as shown below.
result <- format(loess.normalize(mat,mat[15337:15427,1:4], epsilon = 10^-2, maxit = 1, log.it =TRUE, verbose = TRUE, span = 2/3, family.loess ="symmetric"),digits=2,nsmall=2)
where, mat is the matrix of FPKM values for WT and three different tests,
mat[15337:15427,1:4] are the rows corresponding to ERCC FPKM.
I have two questions: 1) Does loess.normalize normalizes other genes of the matrix column wise corresponding to the FPKM values of ERCC for that column or it also takes care of fold change 2)How to provide fold change information, if it doesn't consider fold change
Thanks for any help.
Not sure about the answer, but maybe you can find a hint at https://www.bioconductor.org/packages/erccdashboard ?