What it does
(di)ChIPmunk detects over-represented non-overlapping motifs in fasta sequences.
Which ChipMunk should you choose : Mononucleotides Vs Dinucleotides
Dinucleotide version is better suited to produce a more precise representation of the optimal TFBS binding model. This would allow to properly estimate the number of sequences containing motif hits. e.g. to measure the percentage of “the most reliable” ChIP-Seq peaks in a given dataset.
In terms of the consensus sequence, in general you should get very similar results from the mono- and dinucleotide versions.
Type of the sequence set
Simple : for simple mutil-fasta to be searched in a double-strand DNA mode (the most common choice):
> header1 ACTGTGTGAAA > header2 AGTGTGTGTGTG
You can omit fasta headers since ChIPMunk would simply skip them.
Peak : for peak data with the positional prefences profile (often provided in wiggle-files, .wig). The profile of each sequence should be places in the fasta-header like:
> 1.0 2.0 3.0 2.0 1.5 2.0 AGTAAC > 1.0 2.0 3.0 2.0 1.5 CAGTA
See "Peak multi-fasta generator" in the tool pannel, if you wish to generate peak data.
NOTE that When base coverage information is available, it is highly recommaned to use peak data. This is extremely important for ChIPMunk performance.
If you want to cite ChIPMunk in your research please refer to  for the basic mononucleotide version and to  for the dinucleotide version :
 Deep and wide digging for binding motifs in ChIP-Seq data. Kulakovskiy IV, Boeva VA, Favorov AV,Makeev VJ. Bioinformatics. 2010 Oct 15;26(20):2622-3. doi: 10.1093/bioinformatics/btq488. Epub 2010 Aug24.
 From binding motifs in ChIP-Seq data to improved models of transcription factor binding sites.Kulakovskiy I, Levitsky V, Oshchepkov D, Bryzgalov L, Vorontsov I, Makeev V. J Bioinform Comput Biol.2013 Feb;11(1):1340004. doi: 10.1142/S0219720013400040. Epub 2013 Jan 16.