We will need a clinical correlates table
Closed this issue · 13 comments
Like this one in Rossetti et al. You might want to email David G to help with this as I'm sure he's made many of these.
"Gjertson, David" dgjertson@mednet.ucla.edu
What output should the table be in? I could probably make something in excel pretty easily, but that doesn't seem like an ideal format.
Excel would be great, as we could easily convert it to other things from there.
Do you know how to calculate the p value? It seems like a Fisher test, but I'm not quite sure how to do that for 1 by 1 tables
You really, really should just email Prof. G and work with him on this. He's a professor of biostatistics, and funded on the grant. This could very well take 2 minutes of his time.
Okay, will do.
I asked Dr. Gjertson and he let me know which methods to use, so I should be set for that. However, I don't have any metadata for cohort 3. Is that okay?
Also, should I create a table for all categories/variables here: https://www.pnas.org/content/pnas/116/40/20087.full.pdf or a subset?
I believe he should have metadata on cohort 3.
He can probably make a table of everything like here.
Oh, should I have asked him to make the table? I only asked for the methods/tests to use on the data.
When he gets back to you, you might ask whether it's easier for him to put it together, given I think he makes these all the time.
I asked him about cohort 3. He said that since we are blinded, we don't have clinical data for any of it. He did say that he could make the table, but said he needed the patient IDs and factors we are using for our analysis. I'm not quite sure what he means by the second part. What would you like me to do here?
Ok, let's wait until we know whether some of cohort 3 will be unblinded.
Being handled by Duke.