Downstream workflows assessed
| Approach | Workflow | Description |
|---|---|---|
| BWA-MEM | BWA-MEM | Standard linear reference genome (GRCh38) |
| VG Giraffe | Giraffe(linear) | Standard linear reference genome (GRCh38) with no call set |
| Giraffe(HPRC_pangenome) | HPRC v1.1 frequency-filtered (GRCh38) pangenome including 44 sample assemblies + CHM13 haplotype | |
| Giraffe(1 kGP_pangenome) | GRCh38 phase3 1 kGP call set pangenome | |
| Giraffe(Imputefirst_c5) | Personalized diploid reference using Bowtie 2 + BCFtools + Beagle with HGSVC3 panel at 5× coverage (BBBC5) | |
| Giraffe(Imputefirst_c20) | Personalized diploid reference using Bowtie 2 + BCFtools + Beagle with HGSVC3 panel at 20× coverage (BBBC20) | |
| Giraffe(diploid) | Personalized pangenome using Giraffe's diploid-sampling | |
| Giraffe(benchmark) | Diploid reference created using GIAB truth set | |
| BWA-MEM + LevioSAM2 | Leviosam2(Imputefirst_c5) | Personalized diploid reference using Bowtie 2 + BCFtools + Beagle with HGSVC3 panel at 5× coverage |
| Leviosam2(Imputefirst_c20) | Personalized diploid reference using Bowtie 2 + BCFtools + Beagle with HGSVC3 panel at 20× coverage | |
| Leviosam2(benchmark) | Diploid reference created using GIAB truth set |
[i] For analyses using the HPRC pangenome and 1 kGP pangenome, the sample under analysis and its family members were excluded, if present. The benchmark represents results using the GIAB ground-truth call set for the respective samples, indicating the best achievable performance with each workflow. Giraffe(diploid_reported) refers to variant-calling performance statistics directly reported by Sirén et al. (2024) for the corresponding samples.