Global Median Normalization

Question

[bioinfonerd]

Hi everyone,

I am new to the normalization techniques and would like to know what is Global median normalization and the steps to apply it on Nanostring gene expression dataset.

Thanks!!

[Kevin Blighe]

Hello, have you searched for the answer online?

Answer 1

[Kevin Blighe]

I just noticed your mention of the word 'NanoString'. I processed NanoString data a few years ago and made use of the NanoStringQCPro package in R. Assuming you have your RCC files in a directory called RCCfiles, and that you also have a RLF file, you can read in and normalise your data like this:

library(NanoStringQCPro)

rccSet <- newRccSet(rccFiles=dir("RCCfiles/", full.names=TRUE),
  blankLabel="water",
  rlf="RLFfile/LMF_C3533.rlf",
  experimentData.name="Kevin Blighe",
  experimentData.lab="Department",
  experimentData.contact="[e-mail]",
  experimentData.title="NanoString",
  experimentData.abstract="")

Now we can perform background correction, positive control normalisation, and global median normalisation. This should produce log2 data in the end.

rccSet_norm <- preprocRccSet(rccSet=rccSet,
  doPosCtrlNorm=TRUE,
  doBackground=TRUE,
  doContentNorm=TRUE,
  normMethod="global",
  normSummaryFunction="median")

As above, but perform median normalisation with just housekeeper genes. Housekeeper genes on my panel were:

• GAPDH
• ACTB
• GUSB
• RPL19
• PGK1
• TUBB
• HPRT1

.

rccSet_norm <- preprocRccSet(rccSet=rccSet,
  doPosCtrlNorm=TRUE,
  doBackground=TRUE,
  doContentNorm=TRUE,
  normMethod="housekeeping",
  normSummaryFunction="median")

The log2 normalised counts can then be accessed via:

exprs(rccSet)

These parameters are highly configurable and dependent on the exact experiment set-up that you have. So, please, if you have more questions, then just access the manual from the command line via:

?preprocRccSet

Kevin

[bioinfonerd]

Thanks Kevin, that's really helpful!