I'm having some problems when I calculated PRS by PRsice

Question

[dong]

The error is "Phenotype is a binary phenotype 0 control(s) 3826 case(s) There are no control samples"

And I also notice the 3826 cases is different the actual cases

I have checked my target file(plink),it seems no problem,so I have no idea how to solve this problem

I will appreciate if you have some ideas Thanks very much!

Answer 1

[Sam]

You will need to provide a full log.

I'd assume that your phenotype file contains both the FID and IID and that you have family data. PRSice only perform analysis in founder samples as we did not implement methods that account for relatedness structure. As a result of that, only 3826 of your samples were included in the subsequent analysis, which also happened to all be cases.

[dong]

Thanks for your reply. Yeah,my phenotype data is based on family data(one affected children,unaffeced sibings and normal parents ). I want to calculate PRS of each person in my family data and compare PRS between affeced children and normal sibings or parents

[Sam]

If you know the p-value threshold, you can do --no-regress --bar-levels X --fastscore where X is your p-value threshold and you will get the PRS of all samples. If you don't know the p-value threshold, it is a bit more complicated, but you can still do --no-regress --fastscore and that should give you scores at a few predefined thresholds which you can use for downstream analyses.

[dong]

I will try.Thank you very much!