This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Bsseq problem error at least two dimensions

hello,everyone. i am a new learner on bioinformatics. recently, i have use the bsseq software to deal with the data. but i have meet some error message. which i can not solve. here is my code:

library(bsseq)
setwd("/home/dnamethylation/testbsseq")
data <- list.files(path="/home/dnamethylation/testbsseq", pattern = "*.cov")
bismarkbsseq <- read.bismark(files = data, sampleNames = gsub(".cov", "", data), rmZeroCov = TRUE, strandCollapse = FALSE)
bsseq_bsmooth <- BSmooth(bismarkbsseq, verbose = TRUE, mc.cores = 24)
BS.cov <- getCoverage(bsseq_bsmooth)
keepLoci.ex <- which(rowSums(BS.cov[, bismarkbsseq$Type == "SSA", drop=F] >= 2) >= 2 & rowSums(BS.cov[, bismarkbsseq$Type == "normal", drop=F] >= 2) >= 2) # (here i do not understand why bismarkbsseq$Type == "SSA" , the "SSA" represent what meaning? can i name it "cancer" or other name?)

after that, i got an error message 

Error in rowSums(BS.cov[, bismarkbsseq$Type == "SSA"] >= 2) : 
  'x' must be an array of at least two dimensions

I really do not know how to deal with it? would you please give me some advice? thanks in advance! best wishes xia

software error
• 2,342 views • link
written by 897776177 • 897776177

You should add some colData if you're going to try and use it to filter by.

• 0 views • link
written by Devon Ryan Devon Ryan

thank you so much for your reply.do you mean i should change the command like this? bismarkbsseq <- read.bismark(files = data, sampleNames = gsub(".cov", "", data), colData=c("a1","a2","b1","b2"), rmZeroCov = TRUE, strandCollapse = FALSE)

but when i change the command, it gives me such error message:(colData = c("a1", "b1", "a1", "b2")) is useless

in this step, i store the 4 samples in a directory,should i sepreate them into two directory?how can i add some colData?would you please give me some guidence? thanks in advance xia

• 0 views • link
written by 897776177 • 897776177

You might have to give it to the BSmooth() command, it's been a while since I've used bsseq.

• 0 views • link
written by Devon Ryan Devon Ryan

thanks for your reply, now i have change my command to : bismarkbsseq5 <- read.bismark(files = data, sampleNames = gsub(".cov", "", c("A1","B1","A2","B2")), rmZeroCov = TRUE, strandCollapse = FALSE), and i hope this may works, but i still do not know "SSA" "normal" represent what meaning, if it represents the colname of the data? can i change my command into this? keepLoci.ex <- which(rowSums(BS.cov[, bismarkbsseq$Type == "A", drop=F] >= 2) >= 2 & rowSums(BS.cov[, bismarkbsseq$Type == "B", drop=F] >= 2) >= 2) i am appreciate for your help, would you please give me some advice? thanks in advance xia

• 0 views • link
written by 897776177 • 897776177

i have try to change the name in BSmooth(), but i can not put them into the command. here is the user_guidence instruction: BSmooth(BSseq, ns = 70, h = 1000, maxGap = 10^8, keep.se = FALSE, BPPARAM = bpparam(), chunkdim = NULL, level = NULL, verbose = getOption("verbose")) ,it seems i can not put the name into it. but i am not sure. thanks for you help!

• 0 views • link
written by 897776177 • 897776177

According to the manual you need to give a dataframe to read.bismark(). SSA and such in that code are meant to be your groups, so you'll need to change it to match what you're using.

• 0 views • link
written by Devon Ryan Devon Ryan

thanks for your help! after i change my command into: bismarkbsseq5 <- read.bismark(files = data, sampleNames = gsub(".cov", "", c("A1","B1","A2","B2")), rmZeroCov = TRUE, strandCollapse = FALSE) bsseq_bsmooth5 <- BSmooth(bismarkbsseq5, verbose = TRUE, mc.cores = 24) BS.cov <- getCoverage(bsseq_bsmooth5) when i put the command: head(BS.cov, n=5), it gives me such message A1 B1 A2 B2 [1,] 1 5 2 4 [2,] 1 0 0 0 [3,] 2 5 1 3 [4,] 1 2 0 1 [5,] 0 3 1 1 and when i use the command:colnames(bsseq_bsmooth5),here is the result "A1" "B1" "A2" "B2", it seems that i add the colname to the data, but when i use the command:keepLoci.ex <- which(rowSums(BS.cov[, bismarkbsseq$Type == "A", drop=F] >= 2) >= 2 & rowSums(BS.cov[, bismarkbsseq$Type == "B", drop=F] >= 2) >= 2) ,it still gives me error message like previously: Error in rowSums(BS.cov[, bismarkbsseq$Type == "A", drop = F] >= 2) : 'x' must be an array of at least two dimensions...i am sorry to bother you, but would you please point out which mistake do i made? thanks in advance xia

• 0 views • link
written by 897776177 • 897776177

sorry, i put the wrong command:keepLoci.ex <- which(rowSums(BS.cov[, bismarkbsseq5$Type == "A", drop=F] >= 2) >= 2 & rowSums(BS.cov[, bismarkbsseq5$Type == "B", drop=F] >= 2) >= 2) Error in rowSums(BS.cov[, bismarkbsseq5$Type == "A", drop = F] >= 2) : 'x' must be an array of at least two dimensions, it still give me error message as previously.

• 0 views • link
written by 897776177 • 897776177

1 answer

hello, again,thanks for your persistent help! it seems it works with me after i name the data,here is all my command,i hope this may help:

library(bsseq)
setwd("/home/dnamethylation/testbsseq")
data <- list.files(path="/home/dnamethylation/testbsseq", pattern = "*.cov")
bismarkbsseq5 <- read.bismark(files = data, sampleNames = gsub(".cov", "", c("A1","B1","A2","B2")), rmZeroCov = TRUE, strandCollapse = FALSE)  #(please remember name the data!!!!!name the data!!!!!)
bsseq_bsmooth <- BSmooth(bismarkbsseq5, verbose = TRUE, mc.cores = 24)
bsseq_data_tstats <- BSmooth.tstat(bsseq_bsmooth5, group1= c("A1","A2"), group2= c("B1", "B2"),estimate.var = "group2")
dmr_list <- dmrFinder(bsseq_data_tstats, verbose = TRUE)
write.csv(dmr_list,"no_cutoff_dmr.csv")

this command may not perfect, but it works on my system, and this problem troubles me for a long time, i hope this may help others.

Devon, thanks for your persistent help!!!thank you so much!! best wishes xia

• 0 views • link
updated by Devon Ryan Devon Ryan • written by 897776177 • 897776177

Log in to answer this question.

More posts like this