-
Notifications
You must be signed in to change notification settings - Fork 10
/
NEWS
90 lines (65 loc) · 2.96 KB
/
NEWS
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
lme4qtl 0.2.2
=============
- Throws a warning when misspecified IDs in `relamt` argument.
Related to issue #20 and SO question
https://stats.stackexchange.com/questions/417160/linear-mixed-effects-model-i-cant-seem-to-avoid-either-convergence-errors-or
lme4qtl 0.2.1
=============
- Merged with #18 and its extension #19.
The new update of Z-matrices by the Khatri-Rao product (proposed by @deruncie)
is more elegant and in accordance with the lme4 implementation;
see https://github.com/variani/lme4qtl/pull/18#issuecomment-510646099.
Now creating models with polygenic slopes is possible;
see code examples (model `m6` and `m6b`) in #18.
lme4qtl 0.1.11
=============
- Fixed issue #1 (https://github.com/variani/lme4qtl/issues/1).
lme4 threw the "error with reference class field ‘Ut’, should be from class “dgCMatrix”",
when lme4qtl decomposed the custom covariance matrix as A = R'R.
That was solved using `as(R, "dgCMatrix")` instead of `as.Matrix(R) in `relfac` functions.
lme4qtl 0.1.10
=============
- Fixed issues:
- #5 (https://github.com/variani/lme4qtl/issues/5): repeated measurements and `relmatGlmer`
lme4qtl 0.1.9
=============
- The first public version hosted at https://github.com/variani/lme4qtl
lme4qtl 0.1.8
=============
- The SVD method is replaced by EVD in `relfac` fucntion.
Now the decompoistion K = LL' works for rank-deficient matrices such as (1|FAM).
The test file `test-relfac.R` is updated with necessary checks.
lme4qtl 0.1.7
=============
* Towards the official release before submitting the article to Bionf.
lme4qtl 0.1.6
=============
* The use of variables given in `relmat` are revised.
Now the cases when when these variables are characters or integers
seems to work OK.
The point is that in the order of data rows (used when passed from formula to matrices).
The example that may fail without this update is http://kbroman.org/lmmlite/assets/lmmlite.html
(the heritability is estimated as `0`).
lme4qtl 0.1.5
=============
* The release prepared for the internal use of collaborators.
Two main functions were tested: `relmatLmer` and `relmatGlmer`.
lme4qtl 0.1.4
=============
* The package was renamed from `solaris` to `lme4qtl`.
* The old code was depreciated and the old R files are moved to `docs/R/old`.
* The package was hadleversed as much as needed to pass `devtools::check()`.
solaris 0.1.3
==================
* Updated `solaris` function to hadle `relmat` parameter better.
Now `step` funtion from `lmerTest` package works.
solaris 0.1.2
==================
* Added `relranef` function. This function extracts the random effect vectors `u`
when relative matrices were introduced for the random effects.
See example code in `ins/examples/function-realranef.R`
solaris 0.1.1
==================
* Added `stepRaef` function. This function takes all random effect in a model
and then follow `step`-like procedure to filter out those significant effect.
Likelihood ratio test (LRT) is used.