The fingerprint of a reaction is computed (by the Java code) after
its participants (i.e. compound IDs and coefficients) and direction.
It is checked overnight by the quality checker script (cron job), which
also fixes it in case a bug introduced inconsistencies in the database.
Please refer also to the
javadocs
for updated documentation on fingerprint generation.
Imagine the simple reaction
Its fingerprint could be something like this:
, being 2222 the internal compound ID of A and so on.
The fingerprint is normalized in order to make it unique:
So the following equations would have exactly the same fingerprint that the example above: