I believe it’s the PGS-trait correlation, not SNP heritability or pgs R2. By that an r of 0.355 becomes an R2 of 0.13 or so