[BioC] Timecourse experiment: wrong experimental design?

[andrea.grilli at ior.it]

Hi to all!
I've to perform a time series analysis with timecourse package, but  
I've got plotting problems and some doubt on experimental design.

I've 2 cell lines (first as parenthal, second as transfectant for a  
specific gene), each with 2 biological replicates and 4 time points,  
for a total of 16 samples.

Main goal is to find genes with different temporal behaviour between  
parenthal vs. transfectant. Everything seems to work properly until  
plotting function, that generate this error:

> plotProfile(MB.2D,type="b", gnames= row.names(Data))
Error in rep.pch[1:nrow(con[!is.na(con[, 1]), ]), 1] <-  
pch[1:nrow(con[!is.na(con[,  :
   subscript out of bounds

and plot is empty but title reports probe name, statistics and rank.


Here how I set mb.long function:

# perform indipendent two-samples analysis
> MB.2D <- mb.long(Data, method = "2", times = 4, reps = size,  
> condition.grp = condition, rep.grp = groups, time.grp = time_exp)

And here results of mb.long (that looks right):
# size matrix
> head(MB.2D$size)
      [,1] [,2]
[1,]    2    2
[2,]    2    2
[3,]    2    2
[4,]    2    2
[5,]    2    2
[6,]    2    2

# condition
> MB.2D$con.group
  [1] "Parenthal"    "Parenthal"    "Parenthal"    "Parenthal"    "Parenthal"
  [6] "Parenthal"    "Parenthal"    "Parenthal"    "Transfectant"  
"Transfectant"
[11] "Transfectant" "Transfectant" "Transfectant" "Transfectant"  
"Transfectant"
[16] "Transfectant"

# replicates
> MB.2D$rep.group
  [1] "t0"  "t0"  "t7"  "t7"  "t14" "t14" "t21" "t21" "t0"  "t0"  "t7"  "t7"
[13] "t14" "t14" "t21" "t21"

# time
> MB.2D$time.group
  [1] 1 1 2 2 3 3 4 4 1 1 2 2 3 3 4 4

Input Data is a txt file as matrix of log2 expression data from  
previous filtering steps with limma, imported with:
Data <- read.table("./Input_file.txt", header = TRUE, row.names = 1)

Is it a plotting problem, or could be also a wrong experimental design  
that couses subsequent error?

Further question, for each probe I get a Hotelling statistics value:  
which is the "significant" threshold (I get values from 2330 to  
0.006)? sorry but I'm more confidential with p-value..

Any support is really appreciated,
thanks in advance,
Andrea


Dr. Andrea Grilli
andrea.grilli at ior.it
phone 051/63.66.756

Laboratory of Experimental Oncology
Rizzoli Orthopaedic Institute
Codivilla Putti Research Center
via di Barbiano 1/10
40136 - Bologna - Italy