This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Input data format for MethylSeekR

Hi Everyone,

I have downloaded a BiSulfite-Seq dataset from Encode, which is only in wig format, with the first few lines as following:

track type=wiggle_0 name="UCSF-UBC.Penis_Foreskin_Keratinocyte_Primary_Cells.Bisulfite-Seq.skin03:methRatio" visibility=full color=20,150,20 altColor=150,20,20 windowingFunction=mean
variableStep chrom=chr1
10469   0.347826086956522
10470   0.347826086956522
10471   0.608695652173913
10472   0.608695652173913
10484   0.88

In the description page in GEO, it mentions that the 2nd column are Methylation proportions.

I would like to read in this data into MethylSeekR, as I wish to identify LMR, FMR and UMRs. So far, after searching around in the internet and the package manual I am unable to find a way to do this. I tried using the readMethylome function, but it mentions that (I am copying and pasting):

If format is set to "text" (default), the argument FileName should refer to a tab-delimited text file in the format: chromosome position T M, where each line stands for a CpG, the position refers to the C of the CpG (on the plus strand), T is the total number of reads (total counts) covering the CpG and M the total number of reads without C to T conversion at the C of the CpG (methylation counts). If format="GRanges", the file is assumed to be a GRanges object, containing T and M as first and second data-value entries, saved in rds format.

Is there a way to get T and M from the wig file ? Or any other way to read in the data to use the package?

Thank you,
Tiplu

methylseekr bisulfite-seq
• 3,319 views • link
updated by Ram Ram • written by tiplud • tiplud

Please provide more information, such as a downloadable example data file and your R code. I think we can built such pipeline.

• 0 views • link
updated by Ram Ram • written by Shicheng Guo Shicheng Guo

1 answer

A wig file won't hold the information that MethylSeekR needs. Perhaps you could get T and C counts by just multiplying by a constant for everything and rounding, but this will be approximate and I don't recall enough of the details in MethylSeekR to know if this will cause problems. I fear that you'll need to remap the fastq files if you can't get a more useful format (wiggle files are nice for visualization but aren't very useful for statistics).

• 0 views • link
written by Devon Ryan Devon Ryan

Thank you ! Since I cannot find raw data, I will try multiplying by a constant and rounding. However, as you said, I fear it will be very approximate, and also I have no information about total read counts at each position, so I am not sure how accurate the analysis will be.

• 1 views • link
written by tiplud • tiplud

Good luck! Please do post back to report how well/poorly that works. I doubt you'll be the last person to run into this issue.

• 1 views • link
written by Devon Ryan Devon Ryan

Log in to answer this question.

More posts like this