nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

t-test within a column

DBiber

I would like to perform a t-test for all rows of data. I am working on microarray data and would like to generate some array wide plots with JMP that include p-values. I have two groups of variables (GSM1876278-81 and GSM1876282-85) and am looking for just a simple t-test between the means of these two groups.

I noticed that JMP has some formula tools available but decoding what the inputs are has me a little confused and needing some help.

Thanks

GSE72987_series_matrix_discussion.jmp

Accepted answers

cwillden

Hi @DBiber;

Welcome to the community! Probably the easiest thing to do would be to generate your own t-tests using column formulas since you will have 43,021 individual tests. You can see how I did this in the attached data set. I broke it out across several columns so that it's easier to follow, but I could have done all of it in a single column.

GSE72987_series_matrix_discussion.jmp

cwillden

Great! Glad I could help. Can you mark the answer as the solution if you’re totally satisfied with it?

All comments

cwillden

Hi @DBiber;

GSE72987_series_matrix_discussion.jmp

DBiber

I was told that the community for JMP was very good at answering discussion questions and you've proven this correct cwilden.

This is what I was looking for and will work wonderfully. I will post up the plot I end up with when I'm done just because I would want to see it if I helped out.

I am a little curious why JMP doesn't have what you showed as a canned function in their formula editor. Seems easy enough to just ask for the two groups of data and leave all the maths buried in the function. Regardless, I am very pleased with this method as I have work showing what was done.

cwillden

Great! Glad I could help. Can you mark the answer as the solution if you’re totally satisfied with it?

Mark_Bailey

(I misunderstood the problem so disregard my original reply!)

cwillden

They are comparing groups of 4 columns to each other. The platform options are probably a no-go because they would have tens of thousands of individual reports