Question: factorFootprints, bindingSites and Profile.segmentation
0
gravatar for Nick Gomez
10 weeks ago by
Nick Gomez0
Nick Gomez0 wrote:

Hey all,

I’ve been using the factorFootprints() from the ATACSeqQC package. First basic question is the bindingSites option. Is this supposed to be a Granges object of the ATAC peaks to search through or a list motif binding sites identified by matchPWM()? I’ve been using matchPWM to find instances of the motif across the genome and then intersecting this list with my ATAC peaks and using the resulting intersection as the binding sites. Is this correct way to use the tool to find the motif instances in my ATAC peaks?

Relating to the same function factorFootprints – the output contains a Profile.segmentation with pos distalabun proximalabun and binding. I was just wondering how each of these categories is calculated. In addition what is the range for binding? I'm assuming the higher number the but I don’t yet have a grasp on what constitutes “binding”.

Thanks for your help,

Nick

ADD COMMENTlink modified 10 weeks ago by Julie Zhu4.0k • written 10 weeks ago by Nick Gomez0

The bindingSites option can accept an GRanges object with binding score by any tools that can predict the binding sites. Please note that the "score" must be a column in the metadata. That means you can use matchPWM or fimo to predict the binding sites.

The output of factorFootprints is very similar to centipede and DNAPOS. I add Profile.segmentation with distalabun and proximalabun to present the distal abundance and proximal abundance (which is show as red dash line in the figure). The abundance are calculated by average the signal from the center of binding sites to the end of distal sites you defined, and then use optimal segmentation approach to split then into proximal and distal. The binding is the site motif binding. For more information, you can refer: Epigenome characterization at single base-pair resolution and Genome-wide footprinting: ready for prime time?.

Hope this help.

Jianhong.

ADD REPLYlink written 9 weeks ago by Ou, Jianhong1.1k
Answer: factorFootprints, bindingSites and Profile.segmentation
0
gravatar for Julie Zhu
10 weeks ago by
Julie Zhu4.0k
United States
Julie Zhu4.0k wrote:

Nick,

Thanks for the great questions!

Parameter bindingSites in ATACseqQC can be used to specify the GRanges containing the candidate binding sites, which can be obtained by using matchPWM. However, if you do not have a list of known binding sites already, you can just input pfm for the motif of your interest. The function factorFootprints will fetch the genome-wide binding sites for the input pfm with the specified genome, min.score, and seqlev.

Regarding the implementation and profile.segmenation calculation, please type factorFootprints in a R session. About binding strength, it will depend on the factors. I recommend looking at the binding strength on the binding sites relative to those adjacent to the binding sites. The discussion of footprints at https://www.ncbi.nlm.nih.gov/pubmed/29490630 might be helpful to you.

Best regards,

Julie

ADD COMMENTlink written 10 weeks ago by Julie Zhu4.0k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 16.09
Traffic: 319 users visited in the last hour