Question: EDGE microarray time series analysis fit_models() ERROR
0
gravatar for tleona3
2.9 years ago by
tleona310
tleona310 wrote:

Hi All,

I'm using the bioconductor EDGE package for R to look for differentially expressed genes in a microarray time series dataset. I have a question regarding the parameters on the de_object creation using build_study {edge}.

1) How is the ind argument incorporated into the full/null model creation and fit_models() argument? When I try to add this into my expression set it creates a de_object, but I get an error, please see code below:

#setup data  
eset <- list()
eset$expr <- as.matrix(P20)
eset$ind <- as.factor(P20annot$patient)
eset$time <- as.numeric(P20annot$time)
eset$gender <- as.factor(P20annot$gender)
summary(eset)
Length Class  Mode   
expr   923967 -none- numeric
ind        51 factor numeric
time       51 -none- numeric
gender     51 factor numeric
de_obj <- build_study(data = eset$expr, tme = eset$time, adj.var = eset$gender, ind = eset$ind, sampling = "timecourse")
fullModel(de_obj)
~adj.var + ns(tme, df = 2, intercept = FALSE)
<environment: 0x11dfca8>
nullModel(de_obj)
~adj.var
<environment: 0x11dfca808>
full_matrix <- fullMatrix(de_obj)
null_matrix <- nullMatrix(de_obj)
#Fit the model (likelihood ratio test)
ef_obj <- fit_models(de_obj, stat.type = "lrt")
Error in svd(X) : a dimension is zero
de_lrt <- lrt(de_obj, nullDistn = "normal")
Error in svd(X) : a dimension is zero

My design annotation is below. Please note that each individual was only sampled at three of the five time points. I feel this is what is giving me the error, but I'm unsure why. When I run the code without ind as a parameter it runs just fine.

print(annotation)
        value ind time gender
1    M_115_0h 115    0      M
2    M_122_0h 122    0      F
3    M_117_0h 117    0      F
4    M_126_0h 126    0      M
5    M_129_0h 129    0      F
6    M_140_0h 140    0      F
7    M_142_0h 142    0      F
8    M_143_0h 143    0      M
9    M_146_0h 146    0      M
10   M_148_0h 148    0      M
11   M_152_0h 152    0      M
12   M_156_0h 156    0      F
13   M_132_0h 132    0      F
14   M_147_0h 147    0      F
15   M_150_0h 150    0      M
16   M_153_0h 153    0      M
17   M_155_0h 155    0      F
18   M_129_6h 129    6      F
19   M_140_6h 140    6      F
20   M_142_6h 142    6      F
21   M_143_6h 143    6      M
22   M_152_6h 152    6      M
23   M_156_6h 156    6      F
24  M_115_24h 115   24      M
25  M_122_24h 122   24      F
26  M_117_24h 117   24      F
27  M_126_24h 126   24      M
28  M_129_24h 129   24      F
29  M_140_24h 140   24      F
30  M_142_24h 142   24      F
31  M_143_24h 143   24      M
32  M_146_24h 146   24      M
33  M_148_24h 148   24      M
34  M_152_24h 152   24      M
35  M_156_24h 156   24      F
36  M_115_72h 115   72      M
37  M_122_72h 122   72      F
38  M_117_72h 117   72      F
39  M_126_72h 126   72      M
40  M_132_72h 132   72      F
41  M_147_72h 147   72      F
42  M_149_72h 149   72      M
43  M_150_72h 150   72      M
44  M_153_72h 153   72      M
45  M_155_72h 155   72      F
46 M_132_168h 132  168      F
47 M_147_168h 147  168      F
48 M_149_168h 149  168      M
49 M_150_168h 150  168      M
50 M_153_168h 153  168      M
51 M_155_168h 155  168      F

Also, there is an argument: basis.df = 2. How do I know what # to set the degree of freedom to for this?

Thanks!

microarray time series • 791 views
ADD COMMENTlink modified 2.9 years ago • written 2.9 years ago by tleona310
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1433 users visited in the last hour