Algorithms


Substructure search

After 3DMET Release 2.0, substructure search is available. This supplys a simple database search by using SMILES. However, three level of the search can be applied by the difference about database description and search keys made by the editor. The difference among three levels are shown as below.
  • Level 3: Ignore atom chiralities
  • Level 2: [level 3 set] + Ignore charge + Ignore subset definition
  • Level 1: [level 2 set] + Ignore bond types
For the ignoring characters in the SMILES string, the search becomes stricter on the order from the level 3 to level 1. Two results comparing three levels are shown in the Table 1.

Table 1. Examples for the three levels of the substructure search

Query = benzene Query = phenol
Level hit number an example hit number an example
Level 3 565   30  
Level 2 730 4-aminobenzoate (B00134) 317 formonectin (B00192)
Leve 1 815 cyclohexylamine (B00135) 351 cyclohesanone (B00106)

Query = 1,3-butadiene Query = n-octane
Level hit number an example hit number an example
Level 3 142   193  
Level 2 221 retinal (B00095) 222 Undecaprenyl phosphate (B00093)
Leve 1 3322 L-lysine (B00013) 433 Arachidonic acid (B00061)

It must be paied attention that this result has possible false negatives, because all substructure will not be described as the one sub-string in the whole SMILES string, even if by using canonical SMILES.
 

For the other information, please refer the documents section.