Hello,
I need some help to understand why I am receiving the following error when I ask FDR for the linear model. When I use RAW p-value I do not get the same error.
Also, in Features Details Table, Ad.p.Val it seems all significant features are not significants (p>0.05). Why?
Data, metadata, and procedures I followed are attached.
Thanks,
Data_input.csv (20.1 KB)
Just took a quick look at your data and your questions (no time for actually testing):
Data & metadata: you have only 22 samples, but you would like to control so many different covariates. Have you tried to consider only one factor “Severity” to see if it is OK? If so, add one more (choose one with most impact).
Parameters: look at “P-value cutoff” the current choice is “Raw” p values, but you are looking at adjusted p values or FDR for significant features
You can read our 2022 Nature Protocol for more details
Hello, Jeff!
Thank you for your response.
My primary metadata is severity (categorical). I tested only using “age” as covariate. First with RAW p-value, and it is ok:
But when I select FDR, I received the error:
I tried to select no covariates, first with RAW, then with FDR and the same error is received.
Using RAW:
Using FDR:
What could be happening?
Thank you in advance.
I have the same problem, with a similar set of data and set of steps presented above that produces the Unknown Error Occurred , Any advice/ answer/ solution sure would be appreciated.
Thank you for the note. Finally find some time to investigate the issue - it is caused by a typo in the code which will be triggered when no sig features are detected. It is now fixed (public server will be updated later).
For now, you can just increase p-value threshold to make sure at least one significant features detected to get around the issue
