Note: The development for this package has been stopped, because we decided to rename 'mdist' to 'motifcounter' which can be found here: motifcounter. Moreover, the interface has been changed quite a bit and several features were added in 'motifcounter'.

The package implements functions for elucidating the statistical enrichment of known motif (e.g. from TRANSFAC or JASPAR) in a given DNA sequence of interest (e.g. a gene promoter).

To this end, we provide two analytic approximations for computing the distribution of the number of motif hits in a sequence of given length, which are referred to as the compound Poisson approximation and the combinatorial approximation. While the combinatorial approximation achieves advantageous accuracy if relaxed thresholds for calling motif hits are used, it comes at the cost of an algorithm whose runtime depends on the length of the DNA sequence that needs to be scanned. On the other hand, the compound Poisson approximation achieves similarly accurate results compared to the combinatorial approximation when stringent thresholds for calling motif hits are used, however, using an more efficient algorithm.

Both methods address three important aspects about motif hit enrichment analysis:

• Firstly, it allows for using sophisticated higher-order Markov models as background models, which are more adequate for capturing the content of DNA sequence (e.g. of promoters, enhancers or CpG islands) compared to order-zero background models. Consequently using higher-order background models over order-zero background models might reduce false positive calls substantially.

• Secondly, the self-overlapping structure of motifs is taken into account which is known to affect the distribution of the number of motif hits. This is especially the case for palindromic motifs (e.g. PPARG) and repetitive motif structures (e.g. SP1).

• Third, the methods take motif hits on both strands of the DNA into account

The mdist package can be installed by cloning the git repository and typing

R CMD INSTALL mdist

or via the devtools package

install.packages("devtools")
library(devtools)
install_github("wkopp/mdist")

The package contains a vignette that explains the main functionality.