however, Eagle still achieved better accuracy than all methods except SHAPEIT2 at sample sizes of N≈50,000 and N≈15,000, with only a 9% (s.e. 7%) increase in switch error rate relative to SHAPEIT2 at N≈50,000: 0.78% (0.05%) for Eagle vs. 0.71% (0.05%) for SHAPEIT2. To confirm these results, we performed a similar benchmark of Eagle and SHAPEIT2 on N≈60,000 GERA samples of more diverse European ancestry34,35 (Online Methods) and observed similar results: 0.94% (0.04%) switch error rate for Eagle vs. 0.83% (0.03%) for SHAPEIT2, a 14% (2%) increase (Supplementary Table 3).
