the use of a haplotype scaffold clearly leads to a increase in R2 especially at lower frequencies. For example, at 0.5% frequency the SHAPEIT2 haplotypes made with an 2.5M scaffold increase R2 by 0.1 compared to the 1000GP Phase 1 set of haplotypes. We also find that using the 1M scaffold produces almost identical imputation performance to the 2.5M scaffold. Running SHAPEIT2 without a scaffold produces results intermediate to those of the scaffolded haplotypes and the 1000GP Phase 1 set of haplotypes.
