Can you get information about variance of microarrays experiments from a MA plot ?

Question

[Antonio R. Franco]

Can you actually get this kind of information from a MA plot ?

The plot suggests that the variance of the individual transcriptomic experiments is high and that a high number of genes in the plot biological replicates show high standard deviations and are out of the range for replicates.

This is referred to a microarray experiment with data already normalized If so.. How ?

Answer 1

[Devon Ryan]

If you color dots by significance you can see those with decent A and large M that are not significant. Those will necessarily show higher variance between replicates (otherwise they'd be significant). However, that's basically the extent that one can asses variance from an MA plot and I think it's generally unwise to use it for such purposes (if you want to assess variance, then look at variance or a PCA).