[dpryan]

No, that's not the output from bowtie2 (if someone told you that, they were wrong), it's the output from bsmooth's methylation extractor (or at least it looks like it). The aligner that you tell bsmooth to use has no effect on the formatting of the methylation file.

[ge2sasag]

ah, ok I understand. So according to the files that I have (I have 3 files like that, 2 treatments and 1 control), what would be better to create the BSseq object? I don´t know who to write the code by only reading the manual

[dpryan]

Honestly, with just 3 samples there's no point in any statistical test, the results would be useless. Don't waste your time with this dataset.

[ge2sasag]

Maybe, but this is the only thing we have and I have to do it anyway...

should be some like this?:

read.bsmooth("C:\Users\Gema\Desktop", Sample1.bed, seqnames = NULL, returnRAw = FALSE, qualityCutoff = 20, rmZeroCov = TRUE, verbose = TRUE)

[dpryan]

I'll also add that after a bit of checking/memory refreshing, the format you posted isn't from bsmooth but is similar to bedGraph format, just with the strand column removed. Adding a strand (e.g., awk 'BEGIN{OFS="\t"}{print $1,$2,$3,".",$4,$5}' somefile > somefile.bedGraph) should allow you to use bedGraph2BSseq.py from bison. The instructions for loading the resulting files into R are in the README.md file.

[ge2sasag]

but this command auk is not for R right?

[dpryan]

Correct, awk is run from the command line.

[ge2sasag]

which would be the equivalent in R?

[dpryan]

Well, you end up just reading the file into a dataframe, adding a column of "." and writing to a new file in a different order:

d <- read.table("somefile", header=F)
d$strand = "."
write.table(d[,c(1,2,5,3,4)], file="somefile.new", row.names=F, colnames=F)

Or something like that. Having said that, perhaps you're lucky and the positions covered in each of the three files are the same. Then you just skip all of that with something like:

library(bsseq)
files<- c("sample1.file", "sample2.file", "sample3.file")
fl <- lapply(files, function(x) read.delim(x, header=F))
gr <- GRanges(seqnames=Rle(fl[[1]][,1]), ranges=IRanges(start=fl[[1]][,2], end=fl[[1]][,3]), strand=Rle("*", nrow(fl[[1]])))
M <- sapply(fl, function(x) x$[,3])
Cov <- sapply(fl, function(x) x$[,4])
groups = data.frame(row.names=files, condition=c("WT","Mut","Mut"))
BS <- BSseq(M=M, Cov=Cov, gr=gr, pData=groups, samplesNames=samples)

or something like that. I've not tested any of that, but it should give you the idea.

Given your apparent newness to R, you might want to take a local course before continuing.

[ge2sasag]

thank you very much for your help, I´m just starting with R but right now I´m not able to take any course, I will do it during Christmas vacation when I have more time... but I´m pushed to finish this before...

thank you very much again for stealing your time

[ge2sasag]

I run without errors the code that you wrote me for inserting the column but the output is this:

chr1 564495 1 564496 1
chr1 564501 1 564502 1
chr1 565013 1 565014 1
chr1 565040 1 565041 1
chr1 565262 8 565263 0
chr1 565397 5 565398 1
chr1 565469 2 565470 2

the columns are not tabulated (all the data in the same column), the "." column is not there and the others are not sorted

[dpryan]

The last line should instead be:

write.table(d[,c(1,2,3,6,4,5)], file="somefile.new", row.names=F, colnames=F)

Mea culpa

[ge2sasag]

let´s try

[ge2sasag]

perfect! big thanks! I´ll try the bsseq object now

[ge2sasag]

got this:

> M <- sapply(fl, function(x) x$[,4])
Error: unexpected '[' in "M <- sapply(fl, function(x) x$["