dmr problem

[Magalie L]

Hello,

I am searching for dmr between two plant samples, I’ve already used methykit to identify some DMR but I would like to avoid using sliding windows because the boundaries are not precise enough.

This is why methpipe looks like a good solution for me.

I’ve only used the diffmeth tool and then the dmr tool (version 3.3.1) , because I already have C levels computed for each position, but the output for dmr is quite strange and does not give any DMR.

Here is my command line :


dmr AB_CpG_test.methdiff A_CpG_test.meth B_CpG_test.meth ABCpG BACpG –v

the output for ABCpG:

Chr1 90 0 X:0 0 +
Chr1 91 0 X:18446744073709551615 0 +
Chr1 261 0 X:18446744073709551614 0 +
Chr1 262 0 X:18446744073709551613 0 +
Chr1 371 0 X:18446744073709551612 0 +
Chr1 372 0 X:18446744073709551611 0 +
Chr1 439 0 X:18446744073709551610 0 +
Chr1 440 0 X:18446744073709551609 0 +
Chr1 613 0 X:18446744073709551608 0 +
Chr1 614 0 X:18446744073709551607 0 +


This strange output happens regardless of the sample used in my analysis (CpG or not, different chromosome parts etc..). It is strange because my data contains validated DMR between the two lines.

I’ve also tried to use dmr-hdhmm by I got a segmentation fault.

My meth files are quite similar with those provided for Arabidopsis as an example with the pipeline, my median coverage is around 12x and I used only common position between my two samples..

Here the head of A_CpG_test.meth

Chr1 90 + CpG 0.875 16
Chr1 91 - CpG 1 9
Chr1 261 + CpG 1 29
Chr1 262 - CpG 1 23
Chr1 371 + CpG 1 23
Chr1 372 - CpG 0.9655 29
Chr1 439 + CpG 1 18
Chr1 440 - CpG 1 20
Chr1 613 + CpG 0.9545 22

B_CpG_test.meth

Chr1 90 + CpG 0.8929 84
Chr1 91 - CpG 0.9811 53
Chr1 261 + CpG 0.9858 141
Chr1 262 - CpG 0.9495 99
Chr1 371 + CpG 0.9639 83
Chr1 372 - CpG 1 120
Chr1 439 + CpG 0.9747 79
Chr1 440 - CpG 0.9457 92
Chr1 613 + CpG 1 84


Can you help me to understand and solve the problem ?

Thank you very much,

Magalie

[Ben Decato]

Hi Magalie,

The methcounts files look fine: could you send me the head of the methdiff file, as well as the command you used to run it?  dmr-hdmm is part of an experimental package that is not quite ready for release yet.

Thanks very much!

Ben

--
You received this message because you are subscribed to the Google Groups "MethPipe and MethBase Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to methpipe+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

-- 
Benjamin Decato
PhD Student
Computational Biology Section,
Department of Biological Sciences,
University of Southern California
1050 Childs Way, RRI 408A
Los Angeles, CA 90089-2910

[Magalie L]

Here it is :

my methdiff command line (I've ran it several times, with no error messages, including using "reversed" meth data produced with the awk command in the manual, as my data are from plants)

methdiff  -o A_B_CpG.methdiff A_CpG_test.meth B_CpG_test.meth -v

head of my methdiff file :

Chr1  90    + CpG       0.319143        14      2  75   9
Chr1  91    - CpG       0.290179        9       0  52   1
Chr1  261   + CpG       0.439388        29      0  139  2
Chr1  262   - CpG       0.733051        23      0  94   5
Chr1  371   + CpG       0.639958        23      0  80   3
Chr1  372   - CpG       0.0384106       28      1  120  0
Chr1  439   + CpG       0.476184        18      0  77   2
Chr1  440   - CpG       0.714186        20      0  87   5
Chr1  613   + CpG       0.0437868       21      1  84   0
Chr1  614   - CpG       0.65363 21      0       70 3


Thanks a lot for your help.

[Ben Decato]

Hi Magalie,

I just took a look at what you're doing: the methcounts files are only used for generation of the methdiff files, after which you should use the HMR output.  So your call to dmr should look like this:

hmr -o A_CpG_test.hmr A_CpG_test.meth

hmr -o B_CpG_test.hmr B_CpG_test.meth

methdiff  -o A_B_CpG.methdiff A_CpG_test.meth B_CpG_test.meth -v

dmr AB_CpG_test.methdiff A_CpG_test.hmr B_CpG_test.hmr ABCpG BACpG –v

methdiff calculates the probability of the CpG being significantly differentially methylated using fisher's exact test for each CpG based on the coverage and methylation value.  Then the HMR program output is used to get the differential HMRs, and the number of significantly different CpGs inside those differential HMRs is calculated.  Sometimes that number will be zero, so I suggest filtering the output of DMR with some minimum cutoff of the number of significant CpGs (I use 5 or 10).

Let me know if you have any other questions or this didn't solve the problem.

Thanks!!

Ben

--

You received this message because you are subscribed to the Google Groups "MethPipe and MethBase Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to methpipe+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Magalie L]

Oh my,

Currently I am feeling very stupid.

I DID run the hmr line, and it worked fine, I've also read the most updated manual on the github repository several times to check what I was doing et trying to solve the problem, and I don't know why I kept using the wrong files for all my tests...

Thank you so much, now it is working fine and my file looks ok.

Regards,

Magalie

On Wednesday, March 25, 2015 at 4:45:36 PM UTC+1, Magalie L wrote: