Difference between DML lists

[yaaminiv]

There is a discrepancy between the number of DML @sr320 and I obtained. I think it's because there are differences in our R code.

My list: http://gannet.fish.washington.edu/spartina/2018-10-10-project-virginica-oa-Large-Files/2019-03-07-Yaamini-Virginica-Repository/analyses/2018-10-25-MethylKit/2019-03-13-DML-Destrand-5x-Locations.bed
code: https://github.com/epigeneticstoocean/paper-gonad-meth/blob/master/code/04-methylkit.Rmd

Steven's list: https://github.com/sr320/nb-2019/blob/master/C_virginica/analyses/dml52.bed
code: https://github.com/epigeneticstoocean/paper-gonad-meth/blob/master/code/methylkit_sr.Rmd

The first thing I see is a difference in processBismarkAln. I used mincov = 5, while Steven used mincov = 2. Additionally, I used assembly = v3, and Steven used assembly = v001.

• Mine: processBismarkAln(location = analysisFiles, sample.id = sample.IDs, assembly = "v3", read.context = "CpG", mincov = 5, treatment = treatmentSpecification)
• Steven's: processBismarkAln(location = file.list_10, sample.id = list("1","2","3","4","5","6","7","8","9","10"), assembly = "v001", read.context="CpG", mincov=2, treatment = c(0,0,0,0,0,1,1,1,1,1))

Steven also uses this extra filtering step while I do not:

filtered.myobj=filterByCoverage(myobj_10,lo.count=5,lo.perc=NULL, hi.count=100,hi.perc=NULL)

Both of these discrepancies could affect the number of DML and which DML are identified. Which method is "correct"?

[sr320]

Go ahead and set mincov = 2, and use filterbyCoverage - note hi.count is something recent I was playing with and did not necessarily contribute to bed.

[yaaminiv]

Why would I set mincov = 2 if our goal is mincov = 5?

[sr320]

You set coverage in filterByCoverage As this also provides ability to make hi end threshold. Code wise you only have to make object once - this is what takes the longest; then you can easily further tweak with  filterByCoverage. Thus provides more flexibility.

…

On Apr 5, 2019, 9:33 AM -0700, Yaamini Venkataraman ***@***.***>, wrote: Why would I set mincov = 2 if our goal is mincov = 5? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[yaaminiv]

Would assembly differences also affect DML?

[sr320]

no - that is just naming

…

On Apr 5, 2019, 9:52 AM -0700, Yaamini Venkataraman ***@***.***>, wrote: Would assembly differences also affect DML? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[yaaminiv]

I now get 598 DML: https://github.com/fish546-2018/yaamini-virginica/blob/master/analyses/2018-10-25-MethylKit/2018-10-25-Loci-Analysis/2019-04-05-Differentially-Methylated-Loci-Filtered-Destrand-50-Cov5.csv