This is a test version of Biostars. For the public version, visit https://www.biostars.org.
How to deal with cross-hybridizing Affymetrix (or other) probe sets?

Some probe sets on microarrays cross-hybridise (hybridise to several target genes). If all possible targets were included this could skew the analysis in the favour of the cross-hybridising targets, however if only one is selected (at random?) or if the probe sets are left out entirely you might miss something important.

Is there any general concensus on how to deal with this problem?

term-enrichment gene-ontology
• 2,736 views • link
updated by Ram Ram • written by scottwilliamrobinson scottwilliamrobinson

2 answers

I would recommend using custom cdf files. They redefine the probe sets so that it's one to one mapping to the genes. I worked with Affymetrix arrays and use the cdf files here. They keep a very good update of their cdf files as new mRNA information is released.

• 0 views • link
updated by Ram Ram • written by Ctanes • Ctanes

I had read about these recently actually, I think I will try them out. I hadn't noticed before that they supply both gene-centric and transcript-centric CDFs, that's nice.

• 0 views • link
updated by Ram Ram • written by scottwilliamrobinson scottwilliamrobinson

I asked the maintainer of the Cytoscape GO term enrichment plug-in ClueGO how they handle cross-hybridising Affy IDs (as you can input them directly): "For the moment, we allow several affy spots for one entrez id, but only one entrez id for one affy spot. In this last case, we display the gene with the lowest entrez id (oldest annotation)."

• 0 views • link
updated by Ram Ram • written by scottwilliamrobinson scottwilliamrobinson

Log in to answer this question.

More posts like this