How to use?
The search system uses a
well-known chemical structure editor Ketcher (http://ggasoftware.com/opensource/ketcher). Those
familiar with the editor can skip to the next paragraph. There are simple
instructions given below for the rest.
How to define a chemical
structure?
In
the Tools panel select the different
structures and bond types.
Select the required atoms
from the atom toolbar. Some types of atoms are included into the structures by
default (e.g. into the benzene ring), but they can be edited and swapped for
other atoms. If the required atom is not in the list you can input it manually
– double click on the position you want this element to be in and write its
name into the form that appears.
When the simple selection mode is
active, you can enter in-place editing if click and hold on an atom. This mode
supports the direct input of an atom label and charge, for example: N-, C2+, Na.
You can change atom label simply typing
it when the atom is under cursor and highlighted.
When you double click on an atom, the
properties dialog will appear.
When you double click on a bond, the
properties dialog will appear.
If you click on an up bond and if the up
bond mode is selected, this bond will be flipped. The same is true for down bonds.
How to perform similarity
search by molecule fragments?
The
Skoal search systems allows
to find chemical structures which contain the molecular structures drawn in the
editor, or structures similar to them.
There
are two main similarity search criteria in the system:
The first method compares
molecules by directly comparing their structures, represented as graphs.
The second method involves
the calculation of molecular structure fingerprints by finding specific
features of the structure and encoding them in a bit string.
After creating the structure
you want to find in the editor press the “Search” button. After a few seconds
you will see the search results which show how many chemical structures were
found in the database, their chemical formula, structure, name and the value of
distance metric which describes the similarity of the current structure to the
given search structure.
Different search queries can
return different results for the same input structures. This is caused by the
use of randomized algorithms and the fact that the system only returns the
first 20 results.