Issue with PRSice-v2 output

[JessG1981]

Hi,

I have run my data through PRSice-v2 to generate polygenic indices and have an issue with the output for the best-fit model. It looks like this:

PRS.R2: 0.195985 Full.R2: 0.867389 Null.R2: 0.835064

To me this doesn't make sense, as I thought that Full.R2 = PRS.R2 + Null.R2, but in these figures, the PRS.R2 + Null.R2 = 0.195985 + 0.835064 = 1.031049...

Please can anyone help explain what could be wrong with this?

This is the code that I used:

--a1 A2 \
--a2 A1 \
--bar-levels 0.001,0.05,0.1,0.2,0.3,0.4,0.5,1 \
--base GWAS.data \
--beta  \
--binary-target F \
--bp POS \
--chr CHR \
--clump-kb 250kb \
--clump-p 1.000000 \
--clump-r2 0.100000 \
--cov combined_covar.covariate \
--interval 5e-05 \
--ld g1000_eur \
--lower 5e-08 \
--num-auto 22 \
--out PRS_CO \
--pheno pheno.txt \
--pvalue Pval \
--seed 1286160406 \
--snp MarkerName \
--stat Beta \
--target target.data \
--thread 1 \
--upper 0.5

The phenotype is a continuous variable.

Thanks in advance!

Jess

[JessG1981]

PS if it makes any difference - the target data sample size is less than 20 (I know, it is very small)

[choishingwan]

The r2 calculation is

1 - (1 - full) / (1 - null)

Sam

On Thu, Apr 25, 2024 at 2:12 PM JessG1981 @.***> wrote:

PS if it makes any difference - the target data sample size is less than 20 (I know, it is very small)

— Reply to this email directly, view it on GitHub https://github.com/choishingwan/PRSice/issues/356#issuecomment-2077876533, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAJTRYV4QYK4ZOFTS6CSBCDY7FBRDAVCNFSM6AAAAABGZCQ77WVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDANZXHA3TMNJTGM . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[JessG1981]

Thank you very much! That makes sense now.