Post by nielsr on Nov 6, 2017 13:59:23 GMT
Dear Jian,
Could you please explain me how to interpret the results of a GxE analysis in GCTA? My environmental variables is whether you are in a particular group (1) or not (0). The results are:
Source Variance SE
V(G) 0.690112 0.361567
V(GxE) 0.728265 0.499738
V(e) 4.006128 0.359178
Vp 5.424505 0.086915
V(G)/Vp 0.127221 0.066554
V(GxE)/Vp 0.134255 0.092060
Sum of V(G)/Vp 0.261476 0.065878
logL -10543.639
logL0 -10545.489
LRT 3.701
df 1
Pval 2.7189e-02
n 7837
The SNP-based heritability is 26.1%, but how can I interpret the V(G)/Vp of 12.7% and V(GxE)/Vp of 13.4%. Do these numbers mean that 12.7% of the variance in the outcome is explained by a set of SNPs that have an influence in both groups. And that 13.4% of the variance is explained by a set of SNPs that have a different influence in the two groups? If so, what does "different" mean? I am puzzled, because when I perform GCTA analyis in two samples stratified by the enviromental variable I get the following results:
Subset of individuals in environment 0:
Source Variance SE
V(G) 1.303181 0.563001
V(e) 4.735540 0.564978
Vp 6.038721 0.134820
V(G)/Vp 0.215804 0.092849
logL -5641.741
logL0 -5644.497
LRT 5.512
df 1
Pval 9.4408e-03
n 4032
Subset of individuals in environment 1:
Source Variance SE
V(G) 1.489194 0.459804
V(e) 3.301972 0.456394
Vp 4.791166 0.109212
V(G)/Vp 0.310821 0.095165
logL -4977.242
logL0 -4982.909
LRT 11.334
df 1
Pval 3.8046e-04
n 3880
The SNP-based heritability of 26.1% in the GxE analyis seems to be the average of the estimates obtained in the two stratified samples. However, how do the results in the stratified samples correspond to the V(GxE)/Vp estimate of 13.4%?
Thanks much in advance, Niels
Could you please explain me how to interpret the results of a GxE analysis in GCTA? My environmental variables is whether you are in a particular group (1) or not (0). The results are:
Source Variance SE
V(G) 0.690112 0.361567
V(GxE) 0.728265 0.499738
V(e) 4.006128 0.359178
Vp 5.424505 0.086915
V(G)/Vp 0.127221 0.066554
V(GxE)/Vp 0.134255 0.092060
Sum of V(G)/Vp 0.261476 0.065878
logL -10543.639
logL0 -10545.489
LRT 3.701
df 1
Pval 2.7189e-02
n 7837
The SNP-based heritability is 26.1%, but how can I interpret the V(G)/Vp of 12.7% and V(GxE)/Vp of 13.4%. Do these numbers mean that 12.7% of the variance in the outcome is explained by a set of SNPs that have an influence in both groups. And that 13.4% of the variance is explained by a set of SNPs that have a different influence in the two groups? If so, what does "different" mean? I am puzzled, because when I perform GCTA analyis in two samples stratified by the enviromental variable I get the following results:
Subset of individuals in environment 0:
Source Variance SE
V(G) 1.303181 0.563001
V(e) 4.735540 0.564978
Vp 6.038721 0.134820
V(G)/Vp 0.215804 0.092849
logL -5641.741
logL0 -5644.497
LRT 5.512
df 1
Pval 9.4408e-03
n 4032
Subset of individuals in environment 1:
Source Variance SE
V(G) 1.489194 0.459804
V(e) 3.301972 0.456394
Vp 4.791166 0.109212
V(G)/Vp 0.310821 0.095165
logL -4977.242
logL0 -4982.909
LRT 11.334
df 1
Pval 3.8046e-04
n 3880
The SNP-based heritability of 26.1% in the GxE analyis seems to be the average of the estimates obtained in the two stratified samples. However, how do the results in the stratified samples correspond to the V(GxE)/Vp estimate of 13.4%?
Thanks much in advance, Niels