Computational Biology and Language - Ambient Intelligence for Scientific Discovery - page 30

Fig. 4. Chemical Group Vocabulary: The basic chemical groups that form the building
blocks of the amino acids are shown. The chemical group in each cell in the figure
forms one word in the vocabulary. Thus, the size of chemical group vocabulary is 19.
This vocabulary has been studied in the context of secondary structure analysis by
Ganapathiraju et al [17].

Fig. 5. Secondary structure elements in Lysozyme (PDB ID: 1HEW): Three dimen-
sional structure of a protein is composed of smaller units (secondary structure). (A)
The chain can be followed by guide of the rainbow colors. (B) The same view of the
protein as in A is shown, but repeating elements helix (red), sheet (yellow) and turns
and flexible loops (violet) are highlighted.

To study the effect of varying the vocabulary and alphabet on a typical com-
putational biology task, secondary structure analysis and prediction, we first
investigated different units for this task. Secondary structure refers to regular
units of structure that are stabilized by molecular interactions between atoms
within the protein, the most important interaction being the so-called Hydrogen
(H) Bond. There are 7 distinct secondary structures, broadly called helix, sheet,
turn and loop structures. In helix types, the designating secondary structure is
formed due to H-bonds between carbonyl group and amino group of every 3rd,
4th or 5th residues, and these are called 3 10 −

helix re-
spectively. A strand is a unit that shares long range hydrogen-bond interaction

helix, α− helix and π−