Hello, 

I have 2 groups and each group with 3 replicates. Each sample have 50686 genes with RNAseq read counts for each gene. I performed the following function in R to get the normalisation factors

y <- calcNormFactors(y, method = "upperquartile")

I got the following normalisation factors:

group lib.size norm.factors
WTHC1   WHC 43297217    0.9400238
WTHC2   WHC 46369347    0.9358194
WTHC3   WHC 63202020    1.0778591
WTLR1   WLR 61800948    1.0177353
WTLR2   WLR 66249762    0.9980893
WTLR3   WLR 59552488    1.0382522

But with "TMM" function I got the following normalisation factors 

y <- calcNormFactors(y, method = "TMM")

 group lib.size norm.factors

WTHC1   WHC 43297217    0.9458131
WTHC2   WHC 46369347    0.9676089
WTHC3   WHC 63202020    1.0909037
WTLR1   WLR 61800948    0.9883493
WTLR2   WLR 66249762    0.9905648
WTLR3   WLR 59552488    1.0230927

I do not know which is the better normalisation factor to use? I'm doing the analysis to find the Differential Expressed genes.

Thanks for the help! 

Hi,

which is the "best" normalization is not easy to say, since we don't know the ground truth and in most cases the answer will be dataset dependent. However, there is extensive literature on comparison of normalization approaches, just looking for "RNA-seq normalization" will return many hits in PubMed or Google Scholar. Off the top of my head, a good review / comparison paper is:

https://academic.oup.com/bib/article/14/6/671/189645/A-comprehensive-evaluation-of-normalization

Hope this helps.

Best,

Davide