-
Notifications
You must be signed in to change notification settings - Fork 14
/
TODO
147 lines (80 loc) · 3.48 KB
/
TODO
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
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
March 18
Set prior weights to be counts + prior count
based on Poisson delta method
Nov 27,
note that augmentPriorCount() can cause log CPM values to be different
from standard methods for low counts
Nov 14,
Get BIC for each fit for lmFit(). Alreaday works with lmer
Nov 12, 2023
Push latest version to Bioc
voomWithDreamWeights() should allow users to change pseudocount
Oct 6, 2023
See ~/Downloads/RUSH.RDATA
Since genes are dropped by voomWithDreamWeights(), the out$weights and weightsMatrix are no longer the same size:
https://github.com/GabrielHoffman/variancePartition/blob/3b7b52f14ffe47a28425e05da001fbd785874318/R/voomWithDreamWeights.R#L245C36-L245C47
Need to match names instead
Upload to Bioconductor once remaCor 0.0.16 is accepted to CRAN
June 15, 2023
Parallelize mvTest() with iterator
June 2, 2023
DONE: Reproducable RNG
# Use on.exit({})
# withRandom <- function(expr, seed = 1) {
# old <- .Random.seed
# on.exit({.Random.seed <<- old})
# set.seed(seed)
# expr
# }
# use with options() and omp_set_threads
# Use seed with LS.empirical
March 14, 2023
in deviation() add pearson residuals
need to pass in y= vobj
Feb 28
DONE: Zenith on mvTest results
April 19, 2021
Add treat and topTreat
March 30, 2021
check dream + eBayes with trend and robust
#Authors@R: c(person("Gabriel", "Hoffman", role = c("aut", "cre"), email = "gabriel.hoffman@mssm.edu"))
DONE!
March 18, 2021
info must be casted to a data.frame since DataFrame is not handled correctly
and gives error:
Error in if (inherits(possibleError, "error") && grep("the fixed-effects model matrix is column rank deficient", :
missing value where TRUE/FALSE needed
droplevels in dream
# November 2, 2020
- integrate voomWithQualityWeights() with voomWithDreamWeights
o Need to convert lm.fit() calls to dream()
o Need to dream() to return hatvalues
recommend placing a check to verify if the names in the info matrix, match with the names used in the formula and return an error if not
library(BiocParallel)
# globally specify that all multithreading using bpiterate from BiocParallel
# should use 8 cores
register(SnowParam(8))
# By default it is set to the max number of CPUs
# If you have a very long interactive session, this can crash and you get the error you see
Alternatively, you can specify and restart the parallel backend every function call
fitExtractVarPartModel(..., BPPARAM=SnowParam(8))
fitVarPartModel(..., BPPARAM=SnowParam(8))
dream(..., BPPARAM=SnowParam(8))
# July 18, 2018
Check if there is variation in each gene.
else throw error
# June 16, 2016
Improve warnings in fitVarPart and fitExtractVarPart to show which gene give a warning
# April 16, 2016
Make correlation matrix and linear mixed model equantions in vignette
the same as in supplement. The vignette examples are wrong
Add canCorPairs to the vignette
# Feb 18, 2015
add e_{i,k} in ICC correlation equation
# December 14, 2015
# Add strong warning to fitVarPartModel() that estimates memory usage. In vignette, users should be discoraged from using this interface.
# use refit() to fit the model for each gene:
# cannot do this because refit() doesn't accept new weights
# DE test: show increased power when correlation is overestimated, and decrease false positive rate when correlation is underestimated. (order is correct?)
# adjust doesn't work with varying coefficient model, but it throwings an error now
# run simulations with varying coefficient model