EBSeq HMM non-uniform number of replicates

[Nereful]

Hi there,

We are exploring EBseq HMM for our time-course data analysis. 

Based on the manual, the tool assumes uniform number of replicates through time (I think), given this command:

> CondVector <- rep(paste("t",1:5,sep=""),each=3)

Unfortunately, we have varying number of replicates through time. That is, at 0 hr time point we have 2 reps; at 6 hrs, 3 reps; 12 hrs, 3; 24 hrs, 2. 

How shall we modify this command so we can incorporate this issue?  

Thanks,

Nereful 

[Jan Hapala]

Hello Nereful,

I wish I knew the answer, too.

Actually changing the command you mentioned is not difficult, e.g.:

CondVector <- c("t1", "t1", "t2", "t2", "t3")

in case of two conditions by two replicates and one condition with one replicate (modify accordingly).

The problem is that with uneven number of replicates, the program throws an error. Specifically, the EBSeqHMMTest function returns:

Error in j[[i]] : subscript out of bounds

I would like to know how to run the analysis in such scenario.

Best,

Jan

[Nereful]

Hi Jan,

Not sure whether you were able to figure this out.  

In the CondVector part instead of 

CondVector <- rep(paste("t",1:4,sep=""),each=3)

You can specify each replicate for each time points. 

Hence, modify the command with (example): 

CondVector <- c("h10", "h10", "h1", "h1", "h1", "h5", "h5", "h24", "h24", "h24","h120") 

Note that there is uneven number of replicates across the time-points.

'Hope this helps. 

N

[xinl...@gmail.com]

Hi Jan,

I met the same subscript out of bounds error. Have you solved this problem?

Thanks,

Xin

在 2017年8月24日星期四 UTC+8下午5:46:46，Jan Hapala写道：