Question: would it be a good idea to normalize log2 fold changes with loess?
0
gravatar for tonja.r
3.3 years ago by
tonja.r450
UK
tonja.r450 wrote:

In the Microarray if the data is not center around 0 on MA-plots, it is normalized with LOESS.
Lets say, I have raw log2 fold changes and corresponding means for the genes for two samples. The data is not centered around 0, so would it be applicable to do loess normalization? I am concerned that the gene that shows up-regulation (positive log2 fold change) after the correction could show down-regulation (negative log2 fold change). I do understand that the genes with high log2 fold change (the one that are of the interest) would not be effected that it seems unfair how I might treat other genes with log2 fold changes.

rna-seq chip-seq • 1.7k views
ADD COMMENTlink modified 18 months ago by Biostar ♦♦ 20 • written 3.3 years ago by tonja.r450
0
gravatar for dario.garvan
3.3 years ago by
dario.garvan440
Australia
dario.garvan440 wrote:

Yes, some genes that had a positive fold change will have a negative fold change after normalisation. There is no reason to be concerned that they had a positive fold change before normalisation if their true fold change is negative. Normalisation is supposed to improve your dataset, not leave it the same as it was before batch effects are removed!

The most important consideration about using LOESS is whether or not you expect that there are an approximately equal number of genes that have reduced fold changes as have increased fold changes. In some conditions, such as cancers with high expression of c-Myc, that does not happen, and LOESS should not be used. The best way to normalise a dataset is to put spike-in genes at the same volume in every sample then use the spike-in gene measurements for calculating the correction for each sample.

ADD COMMENTlink modified 3.3 years ago • written 3.3 years ago by dario.garvan440
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1752 users visited in the last hour