Geo And Rank Normalization

Question

[Layla]

I am working with, for example, the disease called acute pancreatitis of unspecified form: GDS: 1731 Control GSM: 84551,84549,84550 For what I know in the GEO database, it appears the GDS file which contains summarized information in the following form:

ID_REF =

VALUE = MAS5-calculated Signal intensity

ID_REF VALUE

1449340_at 390

In one paper it appears that Rank Normalization has been applied to the disease expression data, the question that I have is how to apply this Rank Normalization in the contents of the GDS file? is that necessary? because I think that is has already been threated with the MAS5 algorithm

Thanks

Answer 1

[Neilfws]

Short answer: you are correct, the values contained in the GDS files for this dataset are MAS5 values. They are the only values available for you to use, since the authors have not supplied raw data for you to process.

I only see one paper associated with this dataset and it does not mention rank normalization. However, I believe that this procedure simply involves ordering the raw expression values, then replacing the value with the rank before statistical testing; see for example this paper. My impression is that it is an older normalization method and seldom applied today. Since raw values are not available to you, I guess it cannot be applied to this dataset.