Table 12 Summary of properties of structural alphabets and alphabet designs
From: Protein structure search and local structure characterization
Structural Alphabet | Tung & Yang | de Brevern et al. | Our SA |
|---|---|---|---|
Alphabet Size | 23 | 16 | 18 |
How the alphabet size was determined | Prespecified | Iterative shrinking | BIC |
Clustering | k-means | SOM+HMM | SOM+k-means |
Data Set | Preprocessed (Pair Database) | Preprocessed (PBE-SELECT) | No preprocess (nrPDB) |
Substitution Matrix | BLOSUM-like | BLOSUM-like | Self-Training |
Requires known alignments to build matrix | Yes | Yes | No |
Applicability | Limited | Limited | Modular design More flexible |