MAS 5.0 P/M/A Calls still considered publishable?

Question

[Adamc]

Hi,

I'm looking at something we're putting together based on some Affy microarray data we've had for a while. The original analysis was done based on RMA normalized expression, and considered MAS5.0 P/M/A calls for filtering probesets. I personally don't use MAS5.0 anymore in favor of PM-probe only methods such as fRMA. The P/M/A calls generally agree with qPCR results, so I'm trying to establish if it would be worth recommending that we re-do the calls with a more modern method in order to strengthen perception of the analysis, or if it's not going to matter to reviewers. It's not going to go a bioinformatics-focused journal.

Thoughts?

Answer 1

[Ahill]

Yes, still publishable. While the MAS5 expression measure has taken a beating in the literature for high variance, the MAS5 A/M/P calls do have reasonably good characteristics, and data exists in the literature to back their correspondence to transcript abundance, for example based on latin square spiking experiments. Sure there are more recent methods for making a absent/present call out there. But I've not seen a more recent method that's been shown to be better. If there is one, I'd be very interested to see it. A secondary question might be - do you even need to make A/P calls at all?

[Adamc]

Thanks. They were looking more at expression/absence of expression and not so much at differential expression, so the P/A calls were important.