It is common practice nowadays to assume that the L relevant chemical structure features of the ligand can be encoded as a binary "vector" of 1's and 0's denoting presence (1) and absence (0) of these features i.e., (x_l in {0, 1}) [1}].
Given significant substructure pairs, for an arbitrary compound-protein (graph-sequence) pair, we can compute a binary vector of 10,000 elements where if a significant substructure pair is included, the value of the corresponding element is 1; otherwise zero.
Briefly, the flanking regions of the target genes were amplified using different primer pairs (Additional file 2: Table S5) and ligated into the binary vector of pAg1-H3 conthening thygromycincin resistance gene to form pAg1-H3-nrps4 and pAg1-H3-pks4.
Fu26- GFP, Fu48- GFP: 3xMyc, Fu60- NLS: GFP, and Fu61- GFP: NES, respectively, with Fu76- 2x35S recombined with Fu39-2 through LR reaction (Invitrogen) to make binary vectors of Fu39- 2x35S:: GFP, Fu39- 2X35S:: GFP: 3xMyc, Fu39- 2x35S:: NLS: GFP, Fu39- 2x35S:: GFP: NES.
Finally the sequence-based encoding was concatenated with the contact-based encoding, which produced a binary vector of size 100 + 400 × n.
In each block, say (F = (p_1,p_2,ldots,p_7)), at first we set (V_F = (v_1,ldots,v_7 )) the binary vector of LSB bits (v_1,ldots,v_7) of these elements, that is (v_i= p_i ~mod~ 2) which is considered as an element of (Z_2), for( ~i=1,ldots,7).
For each locus under consideration, the inheritance vector is a binary vector of length 2 n corresponding to the 2 n gametes that gave rise to the individuals.
We considered the vector of p values for these sites, and then considered another binary vector of length 19,170, where an entry is 1 if the corresponding gene is an imprinted gene.
Peptide motifs were classified as disordered or structured by four methods and were encoded as a binary vector of length 2, which resulted in an encoding vector of length 20 k + 2(4).
For ANN training, each of the 20 different amino acids was encoded as a binary vector of length 20.
