Chemistry Reference
In-Depth Information
6.3 QSARS VERSUS THE BLM
In QSARs, the biological activity is also produced by the interaction between the
ligand and the biological receptor of an organism. The only difference is that the
ligand
in QSARs is a chemical species, and the biological receptor is the
biotic
ligand
.
The BLM is very similar to QSARs, in that they both use a series of physico-
chemical and biological parameters that would characterize and describe:
•
the transformation of the ions on their path,
•
the solvation in water of the metals,
•
the chemical speciation,
•
the sedimentation and accumulation in an aqueous environment, and
•
the interaction with the biotic ligand in the organisms present in an aqueous
environment.
However, in a QSAR, these physicochemical and biological parameters are called
descriptors
, and their number is infinitely larger because descriptors of different
nature are used (e.g., physical, chemical, geometrical, quanto-molecular, informa-
tional natures).
The BLM and QSARs are distinct in at least three ways:
1. The BLM correlates the biological activity (toxicity) for a multitude of
metal ions (a set of metal ions) and other chemical species that exist
simultaneously in an aquatic environment, and describes their fate
and their interaction with the biotic ligand (which is equivalent to the
biological receptor in a QSAR) with the use of a multitude of experi-
mental physicochemical parameters such as formation constants, pH, and
temperature.
2. QSARs have a statistical component where each chemical species (ligand)
in a chemical's class is represented by a series of descriptors (the physico-
chemical parameters in the BLM) that are the same for all chemical species
in the class.
3. In QSAR studies, the regression equation that is established gives infor-
mation on the mechanism of interaction, ligand-biological receptor, or
other physicochemical processes that influence biological activity. The
linear regression correlates the descriptors and the biological response as
follows:
A = a
0
+ a
1
X
1
+ a
2
X
2
+….,
where A = biological activity, a
0
, a
1
, … are the regression coefficients, and X
1
, X
2
, …
are the structural descriptors.
The prediction capability of a QSAR depends on the statistical features of the
regression output (regression coefficient r
2
, the a
i
coefficients).
