IPA-Defining the Reference Set-2

The Ingenuity Knowledge Base as the Reference Set

Defining the Ingenuity Knowledge Base as the reference set tells IPA that the molecule list in your dataset file is not sufficient to be considered the complete universe of molecules when conducting statistical calculations for Functional Analysis. In this case, IPA uses the molecules with functions, pathway, or list annotations in the Ingenuity Knowledge Base as the complete reference set.

 

You can also select which portion of the Ingenuity Knowledge Base to use as the reference set.

 

In your experiment, if you assayed:	Then use this as the reference set.
Only genes or proteins	Ingenuity Knowledge Base (Genes only)
Only metabolites	Ingenuity Knowledge Base (Endogenous Chemicals only)
Genes, proteins AND metabolites	Ingenuity Knowledge Base (Genes + Endogenous chemicals)

 

 

Dataset as Reference Set

Defining your dataset file as the reference set tells IPA that the molecules in your dataset file should be considered the complete universe when ranking the statistical significance of functions in Functional Analysis.

 

Using your dataset as the reference set helps to control for experimental bias(bias of Functions and Pathways results toward genes in your input dataset) and literature bias(bias of Functions and Pathways results toward functions and pathways that are far more studied and prevalent in the published literature).  For example, if you have results from a boutique chip or cDNA array, uploading all the genes from the chip into IPA, identifying the subset of significant genes on that chip by using an expression value cutoff, and setting your dataset file as the reference set would allow the application to give a more accurate assessment of which functions are prevalent for your functional analysis genes compared to all genes that were measured in your experiment. Keep in mind however, that your dataset should be sufficiently large (ie. >20,000 identifiers) to be accurately used as a reference set, especially when intending to compare p-value results from different datasets/ analyses.

 

 

Cautionary note for Functions:

When defining the Reference Set, please note that your functional analysis molecules should be a subset of the reference set.  If your functional analysis molecules and the molecules in the reference set are the same group of molecules, then the functions associated with both sets would be the same.  In this case, there would be no statistical significance to functions highly associated with the functional analysis molecule set, since the probability of selecting them out of all functions associated with the reference set would be the same as selecting them by chance alone.  This may cause your Functions Tab to be grayed out in your analysis.

 

Therefore, if all your input molecules are interesting and you wish to consider all of them as Functional Analysis Molecules, you should select the Ingenuity Knowledge Base as your reference set.  If you select your dataset file as the reference set in this case (where either no expression values are assigned to genes or no cutoff is specified), IPA will not return any Functions results. (All results for relevant Canonical Pathways are shown regardless of their degree of statistical signficance.)