The substitution effect database is filled in with known data on glycosylation and phosphorylation effects and contains information on about 150 variants of substitution. The way to find the certain effect is demonstrated on the scheme:
If the desired effect is missing from the database, its engine tries 1.) adding or removing amino group at C2 of the substited residue, 2.) searching for another monomer with the same configuration , e.g. GalNAc instead of FucNAc. If these effects are missing too, database engine tries to vary the type and orientation of the group at C2 of the substituent.
The glycosylation effects for three widespread sugar configurations (glc, gal, man) are represented most completely, so the prediction of structure of polymers built of widespread carbohydrate residues only shows the best accuracy. In the case of not more than one non-carbohydrate or rarely-occurring residue per repeating unit, the probability to find the proper structure among 1st five ones generated is about 95%.