StruMs: Structural Motifs documentation¶
Structural Motifs provide a way to represent DNA motifs based on the shape of the DNA instead of the DNA sequence. These motifs may represent the preferences driving specificity of transcription factors for their binding sites. For more information about the underlying model of StruMs, refer to Introduction to Structural Motifs.
The package described here provides a framework for computing and working with StruMs, based on DNA shape features from the Dinucleotide Property DataBase. Currently available capabilities include:
- Train models from known binding sites
- Use expectation maximization to find de novo motifs
- Provide utility to incorporate additional quantitative features, e.g. DNase hypersensitivity.
- Score matches of the motif to novel sequences