2 replies to this topic

[tcellperson]

I am looking at the data that I received back on my Microarray project and I have this question about the data analysis if anyone can help me understand- Is it possible for a gene to show a fold change and yet be non-significant in terms of p-value ? From what I understand, if a gene exhibits a certain fold change, it has to be significant(with a p-value of less than 0.05). But I am seeing genes on my list which are showing 5 fold or more changed and I am told that they are not necessarily significant!

Can someone please help me if I am missing something?

Thanks!
Ann

[surleen]

tcellperson, on May 5 2010, 10:20 PM, said:

I am looking at the data that I received back on my Microarray project and I have this question about the data analysis if anyone can help me understand- Is it possible for a gene to show a fold change and yet be non-significant in terms of p-value ? From what I understand, if a gene exhibits a certain fold change, it has to be significant(with a p-value of less than 0.05). But I am seeing genes on my list which are showing 5 fold or more changed and I am told that they are not necessarily significant!

Can someone please help me if I am missing something?

Thanks!
Ann

yes this happens especially when u consider low intensity genes. low intensity genes tend to give false positive results therefore people first set cut off >6 for log2 signal intensity of probe sets.

[HomeBrew]

It occurs because, for example, a signal strength changing from 10 to 50 is the same (in fold-change units) as one that changes from 10,000 to 50,000.  In the first case, neither the initial signal nor the final signal reaches the threshold for significance, though the signal strength itself shows a five-fold change.