Hi Enrique,

When processing Gene ST Affymetrix arrrays, a set of probe sets is summarized into a transcipt cluster which is usually associated to a gene. This is commonly done through implementing RMA function (background correction, normalization and summarization). However, as you are pointing out some times there are different transcript clusters that refer to the same gene symbol. If you take a look into NetAffx (Affymetrix web acces database), you will find that these transcript clusters have the same information but different probe sets IDs. Also, the fields  " Transcript Cluster Start" and "Transcript Cluster Length" have slightly different values. Hence, full transcript sequences are not exactly the same. 

In order to avoid duplicated gene symbols, a collapsing procedure could be performed. This is usually done by taking the median of these transcript clusters. This is not mandatory but some tools might have problems with duplicated gene symbols (maybe some pathway analysis tools).

Perhaps, a disadvantage is that you lose track of the sequence and probes for validation could fail. I don't think this would happen in most of the cases.

Let's see what other people say. =)

Hello Enrique,

It is worth to consider one more thing about Affymetrix microarray solution. Affymetrix has it's own method to pick up targets that represents each mRNA on Chip. Sometimes target is more specific, sometimes it is less specific and by that i mean probe that can recognize only one mRNA. Despite this Affymetrix encode this

specificity by using special suffices. That is why you can see multiple targets (probes) for given gene on chip.Code used by Affymetrix looks like this: integerID_(a|at|x|s). Letters is parenthesis encode specifity. One can be less specific, other can be more. Check link below.

http://www.affymetrix.com/support/help/faqs/mouse_430/faq_8.jsp