interpretation of bdgdiff results

[Moshe Olshansky]

Dear List and Tao,

Suppose that I have treatment1 and control1 (t1 and c1) and treatment2 and control2 (t2 and c2) and I run the macs2 bdgdiff command. It produces 3 files: cond1, cond2 and common.

What is exactly in these files?

Does common contain the regions which are enriched in t1 vs c1 (abouve the cutoff) and in t2 vs c2 but not in t1 vs t2 or t2 vs t1? Or is it just those which are enriched in t1 vs c1 and t2 vs c2 irrespectful of the comparison between t1 and t2?

And then what exactly the regions in cond1 are? Is it t1 enriched vs t2 (above cutoff)? Or does it also requie that t1 is enriched over c1?

Thank you,

Moshe.

[Moshe Olshansky]

Following my previous e-mail, a more specific question: when deciding that cond1 is enriched vs cond2, do only t1 and t2 matter, i.e. no use of controls (c1 and c2) is made or they are somehow incorporated as well (and if yes, how)?

Thank you,

Moshe.

--
You received this message because you are subscribed to the Google Groups "MACS announcement" group.
To unsubscribe from this group and stop receiving emails from it, send an email to macs-announcement+unsubscribe@googlegroups.com.
To post to this group, send email to macs-announcement@googlegroups.com.
Visit this group at https://groups.google.com/group/macs-announcement.
For more options, visit https://groups.google.com/d/optout.

[Timothy Parnell]

My understanding (I’m just a user) is that t1, t2, c1, and c2 are all scaled to equivalent depths (hence the need for d1 and d2), enrichment scores are calculated internally between the respective treatment and control for each set, and those enrichments are compared to each other and peaks are called. I don’t know if a log likelihood score is generated first between the two enrichment sets and then peaks called, or if peaks are called individually first and then intersected and likelihood score generated.

In any case, all four bdg files are used. In theory, you could use just two enrichment tracks for each set, but there’s no guarantee that the read depths and enrichment calls would have been equivalent, making the comparison untrustworthy. Using all four and making them equivalent relative depths makes the analysis a little more fair.

Tim

On Aug 25, 2016, at 2:57 AM, Moshe Olshansky <moshe.o...@gmail.com<mailto:moshe.o...@gmail.com>> wrote:

Following my previous e-mail, a more specific question: when deciding that cond1 is enriched vs cond2, do only t1 and t2 matter, i.e. no use of controls (c1 and c2) is made or they are somehow incorporated as well (and if yes, how)?

Thank you,
Moshe.

On Wed, Aug 24, 2016 at 6:40 PM, Moshe Olshansky <moshe.o...@gmail.com<mailto:moshe.o...@gmail.com>> wrote:
Dear List and Tao,

Suppose that I have treatment1 and control1 (t1 and c1) and treatment2 and control2 (t2 and c2) and I run the macs2 bdgdiff command. It produces 3 files: cond1, cond2 and common.
What is exactly in these files?
Does common contain the regions which are enriched in t1 vs c1 (abouve the cutoff) and in t2 vs c2 but not in t1 vs t2 or t2 vs t1? Or is it just those which are enriched in t1 vs c1 and t2 vs c2 irrespectful of the comparison between t1 and t2?
And then what exactly the regions in cond1 are? Is it t1 enriched vs t2 (above cutoff)? Or does it also requie that t1 is enriched over c1?

Thank you,
Moshe.

--
You received this message because you are subscribed to the Google Groups "MACS announcement" group.

To unsubscribe from this group and stop receiving emails from it, send an email to macs-announcem...@googlegroups.com<mailto:macs-announcem...@googlegroups.com>.
To post to this group, send email to macs-ann...@googlegroups.com<mailto:macs-ann...@googlegroups.com>.

Visit this group at https://groups.google.com/group/macs-announcement.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "MACS announcement" group.

To unsubscribe from this group and stop receiving emails from it, send an email to macs-announcem...@googlegroups.com<mailto:macs-announcem...@googlegroups.com>.
To post to this group, send email to macs-ann...@googlegroups.com<mailto:macs-ann...@googlegroups.com>.

[Moshe Olshansky]

Hi Tim,

Thank you for your note.

I understand that the pileup files are scaled. But what happens then? Exactly as you said: are peaks called separately for cond1 and cond2, the non-overlapping peaks become enriched in cond1 and cond2 respectively and the overlapping peaks are merged and cond1 is tested against cond2 (and if so, are c1 and c2 used at this stage as well)? Or is anything else is done?