Calling all bio-stats ppl

8 01 2014

Ok folks, I need help and trying to describe what I need on twitter isn’t really working because of the whole 140 character thing. So has many of you know, I am a drosophilist, which means I work on the best model organism EVAH, unless of course you work on something else but can still help me. cough *not* cough….

Seriously, though I’m going to ask some Qs without giving away any specific details because this isn’t published yet etc. I work on the imaginal wing disc of the drosophila and because flies are the most awesomist organism ever, we use the UAS-GAL4  which allows us to knock down genes in specific regions of the disc, which is really cool because we always have an internal wildtype control. You can do a google search of imaginal disc with uas gal 4 and you’ll find amazing images like these ones

So I use a driver to knockdown geneA and look at its effect. I’m specifically interested in looking at the effects, more specifically if any changes occur in the distrubution of proteinB after RNAi knockdown of geneA. I’ve done this experiment multiple times and imaged alot of discs. I’ve taken z-projections and looked at the mean gray values of ProteinB in the z-projection analysis

I’m not looking at the solid red lines, I’m looking at the red dashes within each box. I’m asking if the mean gray value of the WT box is different than the mean gray value of the RNAi box? IE are there less dashes in RNAi box vs the WT.  I did this analysis in ImageJ and recognize its a crude estimate as there is a lot of white space one both sides, which may mask changes but a girl can’t be picky.

So say  I have done this analysis on 30 discs, each with a WT and RNAi mean gray value (MGV). I’ve used Prism to set up a 1 variable group, with WT in ColumnA and RNAi in ColumnB. I ran a pair-wise t-test but I don’t think that is correct because a pair-wise T-test is asking if the differences between the WT and RNAi is consistent. I don’t care if the differences are consistent because there are many factors which may make 1 disc have a greater difference than another disc. I want to know if the difference is important. If I divide each RNAi MGV by the its WT  by its internal MGV, I get a value that is different than 1. How do I test if this is significant? Do I use a ratio t test? or is any difference significant?

Any ideas on how to do this analysis? Because proteinB is a phospho proteins its staining is very punctate so if you have better idea for how to analysis the images that would be awesome too.





2 responses

9 01 2014
Momma, PhD

If you’re asking if there are less dashes (i.e. the protein staining pattern actually looks like that, or labels some structure with that shape), can you just count the number of dashes in each condition?

9 01 2014

momma PhD – the antibody staining i’m looking at is a bunch of punctate dots. not sure if I can just count them, as that wouldn’t really tell me if the amount has gone up whereas and increase in fluorescent intensity would. And i’d be counting on the order of 1000 pixels per side

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s