Question

Is there any way to set bsseq object as 'smoothed'?

0

Entering edit mode

biohack92 • 0

@biohack92-10159

Last seen 8.6 years ago

I have bsseq objects for that have already been smoothed, and I want to plot the BSmooth methylation estimates using the plotManyRegions function. I had to process each chromosome separately and combine them into a single bsseq object, but I can't use the function as the bsseq object has to have been smoothed by the BSmooth function.

> plotManyRegions(plus, d, extend=2000, addRegions=d)
Error in getMeth(BSseq, type = "smooth") :
  'type=smooth' requires the object to have been smoothed.

Appreciate your help.

bsseq bisulfite • 2.0k views

ADD COMMENT • link 8.6 years ago • updated 8.5 years ago biohack92 • 0

score 0 · Answer 1 · 2016-04-22

Are you sure when you combine data from different chromosomes that it still contains the smoothing estimates? I think my code for combining bsseq objects will drop the smoothing (which in your case it arguably shouldn't, but I think it will). One way to fix this quickly is to just plot the stuff using plotManyRegions on each chromosome. You'll be happy to hear that we are working on more scalable code for this, since we now have experiments with a substantial number of samples, but we don't have anything ready for common usage (and it will be a while before we have). On Fri, Apr 22, 2016 at 12:26 AM, biohack92 [bioc] <noreply@bioconductor.org> wrote: > Activity on a post you are following on support.bioconductor.org > > User biohack92 <https: support.bioconductor.org="" u="" 10159=""/> wrote Question: > Is there any way to set bsseq object as 'smoothed'? > <https: support.bioconductor.org="" p="" 81150=""/>: > > I have bsseq objects for that have already been smoothed, and I want to > plot the BSmooth methylation estimates using the plotManyRegions function. > I had to process each chromosome separately and combine them into a single > bsseq object, but I can't use the function as the bsseq object has to have > been smoothed by the BSmooth function. > > Error in getMeth(BSseq, type = "smooth") : > 'type=smooth' requires the object to have been smoothed. > > Appreciate your help. > > > > ------------------------------ > > Post tags: bsseq, bisulfite > > You may reply via email or visit Is there any way to set bsseq object as 'smoothed'? >

score 0 · Answer 2 · 2016-04-25

0

Entering edit mode

biohack92 • 0

@biohack92-10159

Last seen 8.6 years ago

Hi Kasper,

Thanks for the quick reply. Before combining the individual bsseq objects, all objects have the smoothed designation.

I should clarify that I was able to generate the plots using the plotManyRegions function after splitting by chromosome. Additional filtering post-smoothing reduced the number of DMRs, so rather than sifting through plots in each file/chromosome to grab the plots in the order of ranking of the DMRs, I hoped to combine bsseq objects and use the ranked DMRs for the plotManyRegions.

I'll recheck the script and I hope it's a mistake I made somewhere. Appreciate the update and I look forward to an update!

ADD COMMENT • link 8.5 years ago biohack92 • 0

0

Entering edit mode

No, I don't think you have missed anything. Reading the code, combining smoothed data from different chromosomes will lead to the smooth estimates being discarded. While I was thinking of the perils of combining data which has been smoothed in different ways when I wrote the code, I think your usecase is quite compelling, and I will fix this. But unfortunately not in the next week. Kasper On Mon, Apr 25, 2016 at 10:21 AM, biohack92 [bioc] <noreply@bioconductor.org> wrote: > Activity on a post you are following on support.bioconductor.org > > User biohack92 <https: support.bioconductor.org="" u="" 10159=""/> wrote Answer: > Is there any way to set bsseq object as 'smoothed'? > <https: support.bioconductor.org="" p="" 81150="" #81235="">: > > Hi Kasper, > > Thanks for the quick reply. Before combining the individual bsseq objects, > all objects have the smoothed designation. > > I should clarify that I was able to generate the plots using the > plotManyRegions function after splitting by chromosome. Additional > filtering post-smoothing reduced the number of DMRs, so rather than sifting > through plots in each file/chromosome to grab the plots in the order of > ranking of the DMRs, I hoped to combine bsseq objects and use the ranked > DMRs for the plotManyRegions. > > I'll recheck the script and I hope it's a mistake I made somewhere. > Appreciate the update and I look forward to an update! > > ------------------------------ > > Post tags: bsseq, bisulfite > > You may reply via email or visit > A: Is there any way to set bsseq object as 'smoothed'? >

ADD REPLY • link 8.5 years ago Kasper Daniel Hansen ★ 6.5k