Skip to main content
. 2021 Sep 8;17(9):e1009783. doi: 10.1371/journal.pgen.1009783

Table 1. Columns left to right show: Statistical formulae for GMTE(1), GMTE(0), RGMTE, MR and CAT estimates; Sufficient assumptions each one relies upon to consistently estimate the GMTE (or zero in the case of the GMTE(0) estimate); Estimate-specific confounder test statistics; generic R code to obtain each estimate.

For the GMTE(0) estimate, TCAT=E^[G|T=1]T, for the GMTE(0) estimate T = 1 − T, T* = T G, for the RGMTE estimate T* = TG and T^*=E^[T*|G]. Note that the GMTE(0) estimate does not directly target the GMTE, but rather zero under the PG assumptions.

Estimate Statistical Formula Sufficient Assumptions Confounder Test Fit in R
CAT
β^CAT(Y) E^[Y|T=1]-E^[Y|T=0]E^[G|T=1] {PG1 ∪ PG3} ∩ NUCHom β^CAT(G)=0,β^CAT(Z)=0 Coef. of TCAT in YTCAT + Z
GMTE(0)
β^GMTE(0)(Y) E^[Y|T=0,G=1]-E^[Y|T=0,G=0] PG3 ∩ {(PG1 ∩ PG2) ∪ NUC} β^GMTE(0)(Z)=0 Coef. of T* in YT + T* + Z
GMTE(1)
β^GMTE(1)(Y) E^[Y|T=1,G=1]-E^[Y|T=1,G=0] PG3 ∩ {(PG1 ∩ PG2) ∪ NUC} β^GMTE(1)(Z)=0 Coef. of T* in YT + T* + Z
RGMTE
β^RGMTE(Y) β^GMTE(1)(Y)-β^GMTE(0)(Y) PG1∪NUC β^RGMTE(Z)=0 Coef. of T* in YT+T*+T^*+Z
MR
β^MR(Y) E^[Y|G=1]-E^[Y|G=0]E^[T*|G=1]-E^[T*|G=0] {PG1 ∪ Hom} ∩ {PG2 ∪ NUC} ∩ PG3 β^MR(Z)=0 Coef. of T^* in YT^*+Z