Table 3.

Examples of GP motifs

Examples of GP motifs
Motif	PTr	NTr	PTe	Nte
{MEEIEII :p >= 3}	67	41	67	55
{IQIIIEE : p >= 3}	83	38	92	50
{(I|I)E(E|(I|E)) : p >= 4}	58	37	83	51
{TQ(D|H)(K|C)(D|H)((((D|H)|A)|A)|A)TQ((H|A)|A)TQ((D|H)|A)(I|A) :p >= 7}	33	20	33	23
{M(L|L)CARACAARAA(L|L)RACAA : p >= 6}	8	28	50	44
{AALAALA(A|M)AA.ILAL(A|M)AA(C|M)AV.IL(.|T)A.ILAAALA(.|(A|M))	50	28	25	45
V.ILVAA.ILL(.|T).IA(A|M)AALA(A|M)V.ILV(R|M) : p >= 20}
{(L|(M|A))(L|(L|(M|A)))(L|(M|A))(L|((L|A)|A))M : p >= 5}	83	37	67	54

The table shows examples of the motifs evolved by the genetic programming process targeting the SCOP b.68 fold. In addition to amino acid characters, the motifs are also made from the disjunction operator (|), the wildcard operator (.), and the Hamming distance operator {: p >= x} that specifies the minimum number of characters that must match in the pattern. For each motif, the table shows the relative percentage of sequences matched in the training and test sets. The positive training set has 12 sequences, the negative 3590 sequences. The positive test set also has 12 sequences; the negative set has 226 sequences.