Gene Selection In Different Condition

Question

[Fahmida Yeasmin]

I have a data set containing 2956 genes and normalized their expression values(the experiment is conducted by using affymetrix array).these expression values are the usual summary value calculated for affymetrix array.that means I have only one expression value for one gene and their associated P-value that indicates is there that particular gene is differentially expressed in that sample. and these 2956 genes expression level are given for 39 conditions. How can i detect the number of Important genes?

[Chris Evelo]

Just to be sure we understand correctly what you did. For Affymetrix you typically do quality control first and decide what arrays to use. Then you do normalisation using all good quality arrays and all genes. (You could use arrayanalysis.org to do these steps or to check the documentation of the procedures used). Finally you do statistics to find changed genes. Since you already are down to 2956 genes you must already have done some selection. How did you do that?

[Chris Evelo]

Ehmmm maybe you did not? What species is this about? Were there only 2956 genes on the array?

[Sean Davis]

What is an "important gene"? Without replication (more than one sample for each condition), I think you will find it difficult to determine such "important genes", but perhaps I misunderstand the problem you are trying to solve.

Answer 1

[Pasta]

Your questions is not very clear but from what I understand "important genes" are typically with a P-value <= 0.05 and a fold-change between conditions >= 2. These values are rather popular but adjust them according to your experiments. You also have to consider normalization first but you have to tell us more about your project

[Fahmida Yeasmin]

0

Well Pasta, let me clear my problem. My data set is like as follows:

Affymetrix:CHPSignal   Affymetrix:CHPDetectionPvalue

gene_1 6.2 0.73717

gene_2 31.9 0.51489

By this way 2956 gene's info is given for all the 39 conditions. Say here i have just given one condition's example.same set up is given for all the remaining 38 conditions

Answer 2

[Fahmida]

Well Pasta, let me clear my problem. My data set is like as follows:

    Affymetrix:CHPSignal   Affymetrix:CHPDetectionPvalue

gene_1 6.2 0.73717

gene_2 31.9 0.51489

By this way 2956 gene's info is given for all the 39 conditions. Say here i have just given one condition's example.same set up is given for all the remaining 38 conditions.