Indirect Input Sequence Encoding

The direct sequence encoding methods preserve the order of residues along the sequence string and encode primarily local information. They are, however, not suitable when global sequence features or information content is more important to the application, or when variable length sequences are to be analyzed. This is evident in the intron/exon sequence discrimination (e.g., Uberbacher Mural, 1991 Snyder Stormo, 1993) and protein classification (e.g., Wu et al., 1992 Ferran Ferrara, 1992) problems. 

Other sequence features and information content can be extracted by various scoring mechanisms (e.g., Uberbacher et al1996). Another example of indirect sequence encoding is the computation of hydrophobic moment (Chapter 6.4). 

Encoding Method Window Vector Size Vector or Scalar Value  

Hydropathy Index (of Central AA) ANLAIDV 1 Value varied 

AA = Amino Acid HP = Hydrophobicity, Alph. = Alphabet. Raw data before scaling or normalization. 

---

Input/Output sequence encoding methods (see Table 9.1) PRF direct encoding of residue profile a = a-helix fi = P-sheet C = random coil t = (5-reverse turn L = loop %AA = amino acid composition FEAT = indirect encoding of sequence features MF = memory factor dist = distance Map = Kohonen map (with dimensions). 

---