Table 1 Features calculated for motifs in the metamatti classifier
From: Metamotifs - a generative model for building families of nucleotide position weight matrices
Feature | Description |
|---|---|
Maximum metamotif hit scores with all of the familial metamotifs | Motifs were scanned with all input metamotifs and the optimal score was chosen. Both familially discovered and models discovered. |
Per-column average entropy | Average Shannon entropy of columns. |
MLE Dirichlet parameters | A maximum likelihood Dirichlet distribution is estimated as described in [42] and the parameters of this distribution are used as features (α A ,α G ,α C ,α T ). |
Symmetric Dirichlet background parameters | A symmetric Dirichlet distribution is estimated as described above. |