Methdiff results

[WAN MILES]

Dear methpipe group,

Thanks for your powerful software. I have a question about the methdiff results in the 5th column. Does this value is the p-value of "fisher's exact test"? Value <0.05 means there is significant difference between the two groups? Thanks for your help and looking forward to your reply.

Best
Miles

[Ben Decato]

Hi Miles,

You are correct that this is the p-value outcome of a one-directional Fisher's exact test. It is important to adjust these p-values for multiple testing, so depending on your sequencing depth, you may want to consider aggregating the methylation levels in gene promoters or other genomic features of interest to cut down on the total number of tests. I recommend producing a Volcano plot to show the magnitude difference by p-value, colored by significance after FDR, to make a decision on cutoffs for downstream analysis.

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.
To view this discussion on the web visit https://urldefense.com/v3/__https://groups.google.com/d/msgid/methpipe/6fce2422-d42c-48c3-9671-7a0005b8d035o*40googlegroups.com__;JQ!!LIr3w8kk_Xxm!6diiEgDJbJr4OD9hbP4plEff8QILMJmLnZmDLS-_YmMr-1tHuuHQFsPxRsSJyg$ .

--

-- 
Benjamin Decato, Ph.D.
Computational Biology Section,
Department of Biological Sciences,
University of Southern California
1050 Childs Way, RRI 408M
Los Angeles, CA 90089-2910

[miles...@gmail.com]

Dear Ben,

 

Thanks for your explanation and suggestions. Have a good day!

 

Best regards,
Miles

To view this discussion on the web visit https://groups.google.com/d/msgid/methpipe/CAPVH43VVk4yQ1WbwffDyFe-EWzd_qTCWt2BB5pbR8abgGWBDgg%40mail.gmail.com.

 

[MILES]

Dear Ben,

Hope you are doing well. I met another problem when I got the results from dmr. The output of dmr contains two files and both of them are DMR, right? In the MethPipe manual, it suggests that the results can be further filtered based on 4th and 5th column. Apart from that, is there any more value that can be used to do the filtration/describe the ranking of each DMR row, for example, p value? In fact, I would like to throw the filtered results into IPA software after annotating these DMRs with relevant genes, however, IPA requests a value to describe the ranking of the input genes. 

Hope I described my question clear and looking forward to your reply. Thanks and have a good weekend. 

Best

Miles

On Friday, June 19, 2020 at 11:10:03 PM UTC+8, Ben Decato wrote:

Hi Miles,

You are correct that this is the p-value outcome of a one-directional Fisher's exact test. It is important to adjust these p-values for multiple testing, so depending on your sequencing depth, you may want to consider aggregating the methylation levels in gene promoters or other genomic features of interest to cut down on the total number of tests. I recommend producing a Volcano plot to show the magnitude difference by p-value, colored by significance after FDR, to make a decision on cutoffs for downstream analysis.

Thanks,

Ben

On Fri, Jun 19, 2020 at 12:18 AM WAN MILES <mile...@gmail.com> wrote:

Dear methpipe group,

Thanks for your powerful software. I have a question about the methdiff results in the 5th column. Does this value is the p-value of "fisher's exact test"? Value <0.05 means there is significant difference between the two groups? Thanks for your help and looking forward to your reply.

Best
Miles

--
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 meth...@googlegroups.com.

To view this discussion on the web visit https://urldefense.com/v3/__https://groups.google.com/d/msgid/methpipe/6fce2422-d42c-48c3-9671-7a0005b8d035o*40googlegroups.com__;JQ!!LIr3w8kk_Xxm!6diiEgDJbJr4OD9hbP4plEff8QILMJmLnZmDLS-_YmMr-1tHuuHQFsPxRsSJyg$ .

[Ben Decato]

Hi Miles,

I'd probably make something cool up, and would use the output of methdiff to do it (not dmr) since you have a p-value from the one-directional Fisher's exact test and can directly compute magnitude difference for each CpG from the count columns of the output. The interesting part of this question is that there is a many-to-one relationship between differentially methylated (DM) CpGs and genes.

I would make a discrete definition for a gene promoter (+/- 1kb from the TSS?), since its differential methylation is most likely to be associated with changes in transcription relative to DM of intron/exon CpGs. Then I'd sum the magnitude differences from all the significant DM CpGs in that promoter. This has a built in protection against false enrichment caused by CpGs that are differentially methylated in both directions in the same promoter, since those magnitudes will cancel each other out.

What you'll end up with is high scores for genes with several significantly DM promoter CpGs all changing in the same direction, or a few that change a lot. It's a cool enough score that I would probably plot the histogram and also use it to rank my genes while investigating the biology for the top hits.

Hope this helps,

Ben

To unsubscribe from this group and stop receiving emails from it, send an email to methpipe+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/methpipe/ea0f9aec-ae15-4c21-8644-f90d62e23bd4o%40googlegroups.com.

[miles...@gmail.com]

Dear Ben,

 

Thank you very much for the suggestions. I was thinking about ranking these DMRs based on their methylation rate (M/(M+U) * 100%). But I still need to consider the many-to-one relationship between some DMRs and their corresponding promoter regions. In addition, I thought these DMRs output from hmr were also based on some cutoff methods.

 

Thanks again for your help. I will try the method you suggest. Have a good night!

 

Best regards,
Miles

 

From: Ben Decato
Sent: Monday, July 6, 2020 4:41 AM
To: meth...@googlegroups.com
Subject: Re: [methpipe] Methdiff results

 

Hi Miles,

To view this discussion on the web visit https://groups.google.com/d/msgid/methpipe/CAPVH43Va03wB5d%2BCX5QNxSBp-%3DErp-fWkNkOtsm0bXir4mMZZA%40mail.gmail.com.