Canonical SMILES InChl

The hash origin of InChIKey also means that it is not convertible back to the original InChl or molecular structure, because for each InChIKey there is an unlimited number of possible matching input values. Although this might seem to be a drawback of the format, it is simply the price of the fixed length of the identifier. When a readable identifier with no possible collisions is needed, InChl (or canonical SMILES) should be used. 

SMILES has the additional feature of being human readable, but this is not very important in our model case. InChl, and InChIKey by inheritance, features a much better and robust normalization of structures for example, two different tautomeric forms have the same InChl, but different canonical SMILES. Also, the layered structure of InChl gives us the possibility of excluding some particularity of a structure, such as its stereochemistry, from the search when needed. This is not possible using InChIKey or SMILES. 

Canonicalize chemical structures, i.e., make all chemical structures quickly comparable for a computer. For example, canonical smiles or InChls can be used. 

Both InChl and InChIKey are canonical by default, and there is no other option for them. The situation is different for SMILES, where both a canonical and a noncanoni-cal form exists. For this reason InChl (or InChIKey) is probably better, because the... 

---