Search
Question: help! new to R! Detection p value cut off
0
3.5 years ago by
a0
australia
a0 wrote:

Hi Everyone!

I am so sorry to bother you all. I am new to microarrays/R.

Currently I am looking at gene expression patterns and have done ssn normalization and adjustments. I need to remove all genes that fail p value cut offs. My PI wants to use two different possible cut offs first to determine number of genes significant for either and go from there (P=.05, P=.01). We also want to try and learn look at genes significant under FDR=.1 and FDR=.05 (A previous student had done this for a different expression pattern). Unfortunately noone left in my lab knows how to use R and I cant figure out the code :( Can anyone help me??

modified 3.5 years ago • written 3.5 years ago by a0
1

You don't mention what package is being used in the analysis. We'll be in a better position to help you if we know what's confusing you. At the very least, you'll want to have a look at the user's guide or vignette for that package. That might provide some more clarification regarding the purpose of the code.

ADD REPLYlink modified 3.5 years ago • written 3.5 years ago by Aaron Lun21k
0
3.5 years ago by
a0
australia
a0 wrote:

Hello,

I am using R, bioconductor. I am confused about the user's guide. That's why I am here.I  have gotten as far as sum(GValues<.05) etc. Are those the accurate values? Afterwards based on the previous students notebook they did FDRGValue<-p.adjust(GValues, method="BH") However how they ditermined the number of genes sigfnciant using FDR is confusing. Wouldn't it just be sum(FDRGValues< .05)???? I know I sound like a idiot, that is because when it comes to R I am. I will take a course in it next year but for now I just need help with this so we can identify if we are on the right track with a hypothesis.

1

There are many packages in Bioconductor associated with microarray analysis. For example, are you using affy, limma or something else? How was GValues constructed, and what does it contain? Posting the relevant code would help us give better advice to you.

ADD REPLYlink modified 3.5 years ago • written 3.5 years ago by Aaron Lun21k
0
3.5 years ago by
a0
australia
a0 wrote:

I'm really sorry. I'll talk to the previous grad student or a biostats person in my department. Obviously this is bigger than I ever imagined.

1

No worries. Good luck with it.