|
|
Fitting a Cholesky model (10 genetic and 10 residual parameters) to the data using 4 cores, sharing memory across 24 cores, required 34h using R MKL 3.6.3. Therefore, the model was pre-fitted and saved (`fit.large.RData`) to illustrate `grmsem` follow-up functions. Note that some output has been omitted for clarity.
|
|
|
|
|
|
```{r eval = FALSE}
|
|
|
#Do not run:
|
|
|
#load("G.large")
|
|
|
#load("ph.large")
|
|
|
#fit <- gsem.fit(ph.large, G.large, LogL = TRUE, estSE = TRUE)
|
|
|
|
|
|
> load("fit.large")
|
|
|
> fit
|
|
|
$model.in
|
|
|
part label value freepar
|
|
|
1 a a11 0.01686961 1
|
|
|
2 a a21 0.75022252 1
|
|
|
3 a a31 -0.10352849 1
|
|
|
4 a a41 -0.81445295 1
|
|
|
5 a a22 -0.34964589 1
|
|
|
6 a a32 0.70013464 1
|
|
|
7 a a42 0.30439698 1
|
|
|
8 a a33 0.87503509 1
|
|
|
9 a a43 -0.37434617 1
|
|
|
10 a a44 0.86376688 1
|
|
|
11 e e11 -0.28008894 1
|
|
|
12 e e21 -0.92449828 1
|
|
|
13 e e31 -0.72312602 1
|
|
|
14 e e41 -0.70924504 1
|
|
|
15 e e22 0.51178943 1
|
|
|
16 e e32 0.66207502 1
|
|
|
17 e e42 -0.64331456 1
|
|
|
18 e e33 0.16870994 1
|
|
|
19 e e43 0.30222504 1
|
|
|
20 e e44 0.21833386 1
|
|
|
|
|
|
$model.fit$LL
|
|
|
[1] -4556.647
|
|
|
|
|
|
$model.fit$calls
|
|
|
function gradient
|
|
|
488 118
|
|
|
|
|
|
$model.fit$convergence
|
|
|
[1] 0
|
|
|
|
|
|
$model.fit$message
|
|
|
NULL
|
|
|
|
|
|
$model.out
|
|
|
label estimates gradient se Z p
|
|
|
1 a11 -0.590283919 7.196924e-03 0.018936941 -31.1710286 2.632163e-213
|
|
|
2 a21 -0.406228554 4.760323e-04 0.022331488 -18.1908415 6.099640e-74
|
|
|
3 a31 -0.412282704 -1.306064e-03 0.021811928 -18.9017085 1.104185e-79
|
|
|
4 a41 -0.251181508 4.515862e-03 0.025355870 -9.9062469 3.910610e-23
|
|
|
5 a22 -0.611871764 7.519419e-03 0.013081543 -46.7736703 0.000000e+00
|
|
|
6 a32 -0.412494347 5.020765e-05 0.015153798 -27.2205248 3.712770e-163
|
|
|
7 a42 -0.406496327 2.858898e-03 0.020248170 -20.0757069 1.203583e-89
|
|
|
8 a33 0.443344860 -2.806537e-04 0.013549030 32.7215213 7.719589e-235
|
|
|
9 a43 -0.015668913 1.437494e-03 0.024698450 -0.6344088 5.258141e-01
|
|
|
10 a44 0.655953940 -5.747793e-03 0.014789225 44.3535029 0.000000e+00
|
|
|
11 e11 -0.794244501 2.325624e-02 0.012513836 -63.4693048 0.000000e+00
|
|
|
12 e21 -0.290249172 -8.449568e-04 0.013042715 -22.2537381 1.037729e-109
|
|
|
13 e31 -0.445243952 -5.908758e-03 0.012154714 -36.6313812 9.056960e-294
|
|
|
14 e41 0.006719942 7.527450e-03 0.011697368 0.5744833 5.656408e-01
|
|
|
15 e22 0.521077401 -1.251593e-02 0.008841668 58.9342896 0.000000e+00
|
|
|
16 e32 -0.054613913 1.163049e-03 0.010020443 -5.4502495 5.029922e-08
|
|
|
17 e42 -0.018229268 -8.354153e-03 0.011967918 -1.5231779 1.277142e-01
|
|
|
18 e33 0.405590017 -2.845990e-02 0.007471397 54.2856995 0.000000e+00
|
|
|
19 e43 -0.022769655 7.889148e-04 0.012392180 -1.8374213 6.614772e-02
|
|
|
20 e44 -0.467128074 1.039490e-02 0.009007167 -51.8618176 0.000000e+00
|
|
|
|
|
|
$k
|
|
|
[1] 4
|
|
|
|
|
|
$n
|
|
|
[1] 20000
|
|
|
|
|
|
$n.obs
|
|
|
[1] 5000 5000 5000 5000
|
|
|
|
|
|
$n.ind
|
|
|
[1] 5000
|
|
|
|
|
|
$model
|
|
|
[1] "Cholesky"
|
|
|
|
|
|
$con
|
|
|
[1] FALSE
|
|
|
|
|
|
$ph.nms
|
|
|
[1] "Y1" "Y2" "Y3" "Y4"
|
|
|
|
|
|
attr(,"class")
|
|
|
[1] "gsem.fit"
|
|
|
|
|
|
```
|
|
|
|
|
|
The variance can be obtained as
|
|
|
|
|
|
```{r eval = FALSE}
|
|
|
> var.fit <- gsem.var(fit)
|
|
|
> print(var.fit)
|
|
|
$VA
|
|
|
1 2 3 4
|
|
|
Y1 0.3484351 0.2397902 0.2433639 0.1482684
|
|
|
Y2 0.2397902 0.5394087 0.4198747 0.3507607
|
|
|
Y3 0.2433639 0.4198747 0.5366833 0.2642885
|
|
|
Y4 0.1482684 0.3507607 0.2642885 0.6588525
|
|
|
|
|
|
$VA.se
|
|
|
1 2 3 4
|
|
|
Y1 0.02235634 0.01660915 0.01767332 0.01514868
|
|
|
Y2 0.01660915 0.02017283 0.01672469 0.01562045
|
|
|
Y3 0.01767332 0.01672469 0.02042099 0.01515375
|
|
|
Y4 0.01514868 0.01562045 0.01515375 0.02078249
|
|
|
|
|
|
$VE
|
|
|
1 2 3 4
|
|
|
Y1 0.630824327 0.23052881 0.35363256 -0.005337277
|
|
|
Y2 0.230528809 0.35576624 0.10077361 -0.011449317
|
|
|
Y3 0.353632560 0.10077361 0.36572812 -0.011231587
|
|
|
Y4 -0.005337277 -0.01144932 -0.01123159 0.219104559
|
|
|
|
|
|
$VE.se
|
|
|
1 2 3 4
|
|
|
Y1 0.019878092 0.012210347 0.013679555 0.009287867
|
|
|
Y2 0.012210347 0.012054860 0.009085111 0.007093293
|
|
|
Y3 0.013679555 0.009085111 0.012413525 0.007220553
|
|
|
Y4 0.009287867 0.007093293 0.007220553 0.008358356
|
|
|
|
|
|
$VP
|
|
|
1 2 3 4
|
|
|
Y1 0.9792594 0.4703190 0.5969964 0.1429311
|
|
|
Y2 0.4703190 0.8951749 0.5206483 0.3393114
|
|
|
Y3 0.5969964 0.5206483 0.9024114 0.2530569
|
|
|
Y4 0.1429311 0.3393114 0.2530569 0.8779571
|
|
|
|
|
|
$RG
|
|
|
1 2 3 4
|
|
|
Y1 1.0000000 0.553110 0.5627767 0.3094522
|
|
|
Y2 0.5531100 1.000000 0.7803720 0.5883800
|
|
|
Y3 0.5627767 0.780372 1.0000000 0.4444523
|
|
|
Y4 0.3094522 0.588380 0.4444523 1.0000000
|
|
|
|
|
|
$RG.se
|
|
|
1 2 3 4
|
|
|
Y1 0.00000000 0.02488356 2.233207e-02 3.056075e-02
|
|
|
Y2 0.02488356 0.00000000 1.557784e-02 1.976462e-02
|
|
|
Y3 0.02233207 0.01557784 3.008489e-18 2.206251e-02
|
|
|
Y4 0.03056075 0.01976462 2.206251e-02 4.557191e-18
|
|
|
|
|
|
$RE
|
|
|
1 2 3 4
|
|
|
Y1 1.00000000 0.48661851 0.73623906 -0.01435621
|
|
|
Y2 0.48661851 1.00000000 0.27937355 -0.04100828
|
|
|
Y3 0.73623906 0.27937355 1.00000000 -0.03967676
|
|
|
Y4 -0.01435621 -0.04100828 -0.03967676 1.00000000
|
|
|
|
|
|
$RE.se
|
|
|
1 2 3 4
|
|
|
Y1 1.111452e-17 0.01770789 1.099568e-02 2.499190e-02
|
|
|
Y2 1.770789e-02 0.00000000 2.164591e-02 2.549970e-02
|
|
|
Y3 1.099568e-02 0.02164591 5.397777e-18 2.555311e-02
|
|
|
Y4 2.499190e-02 0.02549970 2.555311e-02 7.971451e-18
|
|
|
|
|
|
``` |
|
|
\ No newline at end of file |