#14 Preprocessing depends on order atoms are defined

[Timo Böhme]

The preprocessing step in standard InChI (1.02) creation which normalizes hydrogens etc. depends on the order in which atoms are defined in input structure. The attached test cases demonstrate this.
The structure in test case 1 (which is a substructure of PubChem CIS 3353305) produces InChI=1S/C4H7N5O2/c5-8-2-3(10)7-1-4(11)9-6/h2,5-7,10H,1H2/b3-2- whereas in test case 2 we get InChI=1S/C4H6N5O2/c5-8-2-3(10)7-1-4(11)9-6/h2,6-7,10H,1H2/q-1/p+1/b3-2- (one hydrogen removed by preprocessing).
Both test cases (SDF files) differ only in the order atoms are defined. In test case 2 the charged Oxygen is moved from last to first position in atom block and bond atom numbers are updated accordingly. Thus both test cases describe the same molecule but get different standard InChIs.

Timo

[Timo Böhme]

test case 1: all hydrogens remain in InChI

[Timo Böhme]

test case 2: preprocessing removes one Hydrogen