1
$\begingroup$

I notice when using tools like HOMER there are two different types of outputs: de novo motifs vs known motifs and it seems no matter what the data is (type of ChIP etc.) there is always a "de novo" motif set in the results.

I was hoping to reach out to experts on insight about how much of the space of motif based structure, within a genome, has been identified. Is there any consensus on what portion of the motif like patterns, for example in human or mouse genome, have been identified? When I find de novo motifs using HOMER in my ChIP-seq or ATAC peaks, what are the chances that in my common-place ChIP experiment (PU.1 for example) I find de novo motifs, not recorded in the databases (i.e. JASPER).

$\endgroup$
1
  • $\begingroup$ Thanks. I previously closed this question mostly due to the "why" in the title. I've reopened the question because it looks more likely to encourage specific answers. $\endgroup$
    – gringer
    Apr 14, 2023 at 0:49

1 Answer 1

2
$\begingroup$

De novo means that the software infers presence of motifs based on the provided sets of DNA sequences, most commonly peaks from ChIP-seq and similar experiments. So essentially it stacks the sequences on top of each other and checks for presence of conserved nucleotides that form sequences of certain lengths, for example < 10bp.

Known means that each peak is checked whether it contains a known motif (so is similar to a given known motif) and then in the end a per-motif statistic is calculated to decide whether enriched or not.

If the investigated protein is well-studied then de novo and known is likely similar or identical. De novo does not mean that found motif is novel in terms or unknown or undescribed. It just means that it is derived from the data directly rather than matching known motifs onto the peaks. It's simply two different ways to approach the question. Hope that makes sense.

$\endgroup$
2
  • $\begingroup$ Thanks @Atpoint, greatly useful and helpful answer. Part of my question remains, albeit in a HOMER independent context, as to how much of the motif space remains undiscovered, if any? $\endgroup$
    – Zebra Fish
    Apr 18, 2023 at 1:03
  • $\begingroup$ Probably quite some. Imagine just for how many transcription factors there are no antibodies available. $\endgroup$
    – ATpoint
    Apr 18, 2023 at 8:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.