1) RPKM means you have already normalized your data. 

2) RPKM has extensively been shown to not be a robust normalization approach (see Wagner, G. P., Kin, K. & Lynch, V. J. Measurement of mRNA abundance using RNA-seq data: RPKM measure is inconsistent among samples. Theory Biosci. 131, 281–285 (2012). ) for more information, as well as other studies that cite this work.

You have not said what you want to do with your data: differential gene expression? Clustering? Something else?

If it's differential gene expression, I would suggest having CLC export your sam/bam files, then using htseq to generate a count table that you can then use edgeR, DESeq2 or limma voom, for example. The latter are all R packages with excellent documentation.

Miriam,

The normalization and normal transformation are very common procedures that are applied to microarray data in order to remove unwanted environmental bias from the experiment. This obviously changes the numeric appearance by scaling the numbers, but this methodology is proven to bring array experiments onto the same scale and make them comparable. Apart from the data transformation and normalization you might want to perform a PCA on your data (can be done in Gene spring).

Normal transformation

Microarray raw data are generally not ready to use for in data analysis and they require a few processing steps of which the normal transformation is an important step. The aim of a normal transformation is that data are transformed into a Gaussian distribution also called normal distribution. This enables the use of statistical methods that assume that data are normal distributed and it helps with the detection of technical glitches in the experiments. I guess there are some data visualization methods in Gene spring that would allow you display single array data and filter out the dodgy one if that is required.

Data normalization

Data normalization is a process that helps to remove environmental bias from experiments and brings all arrays on the same level so that they can be compared. There are different strategies for data normalization and the approach that was selected in this case has been tried and tested on protein array repeat experiments and the array experiments were analysed using mathematical correlation tests. The idea behind this experiment was that a repeat experiment should give near identical results and any deviation from the expected result in the raw data is caused by environmental effects. This data normalization study was performed on numerous occasions using batches of 24 protein arrays and different operators on different days.

The results of this experiments suggested that a data normalization procedure such as quantile, median or rank based normalization procedures that are beneficial for the analysis of expression arrays.

If you have access to R or some decent data visualization software such Matlab or Tableau than you plot and test your data with those.

First of all test if your data are normal distribute – if not then normal transform them (log2 or boxcox …).

Then you can try if normalization improves your data.

Kind regards,

Jens

http://www.biomarker-discovery.com/

1) RPKM means you have already normalized your data. 

2) RPKM has extensively been shown to not be a robust normalization approach (see Wagner, G. P., Kin, K. & Lynch, V. J. Measurement of mRNA abundance using RNA-seq data: RPKM measure is inconsistent among samples. Theory Biosci. 131, 281–285 (2012). ) for more information, as well as other studies that cite this work.

You have not said what you want to do with your data: differential gene expression? Clustering? Something else?

If it's differential gene expression, I would suggest having CLC export your sam/bam files, then using htseq to generate a count table that you can then use edgeR, DESeq2 or limma voom, for example. The latter are all R packages with excellent documentation.

I agree with Darya - RPKM is clearly an inferior normalization scheme for RNA-seq data and the literature on that fact is fairly extensive now.

Most of the current used methods for differential expression analysis begin with raw count data, and most are available freely in R/BioConductor (there is also CuffDiff which is a standalone package).  These tools include their own specific scaling, variance and dispersion estimation and normalization, along with statistical tests appropriate to those normalizations.

I do not know what exactly CLC biobench software includes for RNAseq differential expression analysis, but if it restricts you to RPKM values only, then I would suggest not using it and learn to use the tools in BioConductor instead.

I am not familiar with such data, but I frequently transform the raw data into normal quantiles (or normal scores) so that I get z values that in addition to being perfectly "normal" also assist in the interpretation of the results. 

There is an article that cautions against using such procedures in genetics.

http://link.springer.com/article/10.1007%2Fs10519-009-9281-0

In a nut shell: "one would not use an INT[inverse normal transformation] when the normality assumption is met"

If you expect to have non-normal data, then I believe that such transformations can be very effective in many instances. It depends on the statistical application. 

I agree with Darya and Michael, RPKM is not a good sollution, FPKM would be better. I prefere RSEM normalization:

http://www.biomedcentral.com/1471-2105/12/323

RSEM normalization can be done directly from fastq files, given the genome reference and it will return FPKM and TMM normalized expression values.

Normalized values you can then upload to R (or any other software for secondary analysis).

As for transfromation - do you mean log transformation ? Because that is true that RNAseq data normalized are as far as I am concerned not normally distributed. Log transformation usually is fine for solving that problem. I would suggest to check after that if you have normal distribution by shapiro-wilk test (available in R).

Hope I helped,

Bartosz

I would suggest to go for edgeR based analysis. Where you would also see the data clustering (MDS), so you dont need another multidimensional reduction method like PCA. Anyway, PCA can also be performed. And yes, read count/RPKM none of them are normally distributed and for that depending up on the questions you want to answer several statistical methods are already implemented in edgeR too. If want to answer any thing specific, out side the edgeR domain, you have to mention it clearly including your sample type (Paired or not) etc.