Figure 1: MUSSEL Workflow.

[Step 0] apply LDpred2 to each of the K training populations (ancestry groups) to obtain estimated causal SNP proportions ($p_k$, $k\hspace{0.17em}=\hspace{0.17em}1,\dots ,K$) and heritability ($h_k^2$, $k\hspace{0.17em}=\hspace{0.17em}1,\dots ,K$) parameters based on the tuning set, these parameters will be used to specify the prior distributions and tuning parameter settings for Bayesian learning with MUSS. [Step 1] MUSS: jointly model across all training populations to obtain a total of ($\text{L}\times \text{K}$) PRS models under L different tuning parameter settings for $\text{Pr(}{\delta }_{1j},\dots ,{\delta }_{Kj})$ (functions of $p_k\text{s}$) and ${\rho }_{k_1{k}_2}\text{s}$ across $K$ training populations. [Step 2] for the target population, conduct ensemble learning (EL) via a super learner (SL) algorithm with 3 base learners (elastic net regression, ridge regression, and linear regression) to train an “optimal” linear combination of the ($\text{L}\times \text{K}$) PRS models from the MUSS step to obtain the final MUSSEL model. The prediction performance of the final PRS derived using MUSSEL should be evaluated on an independent testing set.