Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

Semantic Similarity on Genes via Gene Ontology: Relatedness or complementarity?

George Fou gio.fou at gmail.com
Fri Jul 14 03:54:11 PDT 2006


Hi, I am trying to improve classification results on microarray datasets.

I have read in some papers that feature selection is the most common way to reveal the underlying knowledge of a dataset in order to get better results (higher accuracy).

Feature selection is performed by selecting the most highly relative genes (attributes) among all. Some sceptics say that this hides a great amount of risk because selecting the most highly relative genes causes redundancy (because perhaps one possible biological pathway will be supported).
Because I am an Electrical Engineer could anyone explain the above better to me from a biological view? 

Furthermore, I have also read that in any case choosing complementary genes (attributes) among all would increase classification accuracy.

I tried to find similarities among genes exploiting knowledge derived from Gene Ontology. More specifically, I tried to fing semantic similarity between genes based on the terms that those genes are annotated to. Could anyone more connoisseur on this subject tell me if this way gives us highly relative genes or complementary ones?

Could also a biologist tell me a possible way of finding complementary genes exploiting knowledge from Gene Ontology?

Thank you very much for your precious time

Regards,
George
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/go-friends/attachments/20060714/c79191f8/attachment.html>


More information about the go-friends mailing list