Dictionary definition

KDM is based on a dictionary of k-mers of a given length, possibly including a number of undefined positions (gaps, written N). A dictionary can also be defined as double strand, in which case each k-mer and its reverse complement are considered the same entity.

Everything downstream — motifs, sequences, scores — lives in the space this dictionary defines, so it is the first choice you make. To define one, use the function KDMDictionary, which requires the following parameters:

Function: KDMDictionary

RNA-binding proteins

For instance, we can define a dictionary as the one we use for RBP binding proteins as follows:

dictionaryRBP<-KDMDictionary(kmerlength=6,ngaps=2,double.strand=FALSE)

In this case the kmers field will contain:

     kmer
1  NNAAAA
2  NANAAA
3  ANNAAA
4  NAANAA
5  ANANAA
6  AANNAA
7  NAAANA
8  ANAANA
9  AANANA
10 AAANNA
. . .

Transcription factors

For transcription factors we define:

dictionaryTF<-KDMDictionary(kmerlength=8,ngaps=2,double.strand=TRUE)

obtaining:

       kmer  kmer_rc
1  NNAAAAAA TTTTTTNN
2  NANAAAAA TTTTTNTN
3  ANNAAAAA TTTTTNNT
4  NAANAAAA TTTTNTTN
5  ANANAAAA TTTTNTNT
6  AANNAAAA TTTTNNTT
7  NAAANAAA TTTNTTTN
8  ANAANAAA TTTNTTNT
9  AANANAAA TTTNTNTT
10 AAANNAAA TTTNNTTT
. . .

With double.strand = TRUE the dictionary gains a kmer_rc column holding the reverse complement that each entry is pooled with.

The rest of this tutorial uses dictionaryTF, since the worked example is the transcription factor NRF1.