not interested in processes where the matter interacts with light or laser field, it is
pragmatic to use classical mechanics to describe the molecular systems which
involves relatively simple mathematics, i.e. solving Newton’s equation of motion to
describe the interaction within system and their association with other systems and to
model their time evolution and their response to external thermodynamic variables
like temperature and pressure. As per classical mechanics, once we have the
force-field information for a system, its entire future and past can be predicted by
solving equation of motion. Force-fields can be developed by using various available
structure databases and thermodynamics data. In this chapter, We briefly cover
available force-field methods for computing the binding affinity in order to rank
protein–ligand complexes in drug discovery and design. In addition, briefly discuss
their limitations and also present the recent advancement in computational modelling
approaches based on quantum mechanical theory and machine learning algorithm in
a way suitable for drug discovery applications.
3 Free Energies Relevant to Describe Potency
and Pharmacokinetics
Free energy is the key variable that dictates the structure of biomacromolecular
complexes (protein–ligand protein, membrane–ligand, DNA–ligand etc.) and controls various molecular association and ligand transport processes. When there are
many structures possible, the one with least free energy is the most stable one.
Moreover, biomacromolecular or molecular association processes such as drug
binding to receptor, protein–protein binding and drug transport involve the minimization of Gibbs free energy (DG). Any process that involves lowering of Gibbs
free energy can proceed spontaneously. By calculating the free energy change, we
can predict whether an association process is feasible or not. In the case of a drug,
the most relevant aspect is to understand its binding affinity or potency towards a
target biomacromolecule and its association with transport proteins like albumins
[9] and metabolizing enzymes such as cytochrome P450s (CYP) [10–12] and with
glycoproteins responsible for absorption. The ligand binding to target biomacromolecule, transport protein and metabolizing enzymes is dictated by the change in
free energy of the ligand bound to these targets when compared to that in aqueous
solution. Further, it is also necessary to understand whether the compounds will
pass through certain cell membranes and also how well it will dissociate once it is
taken through oral dose which is dependent on the physicochemical properties like
lipophilicity [13] and aqueous solubility [14]. Schematic representation of various
PK computational modeling is shown in Fig. 1, and free energy of relevance is
provided in Table 1.
Computing the free energy of binding of the drug with a biological target and
other targets (such as glycoproteins, albumins, cell membranes) that mediate the
drug transport across the body to relevant target area is the main goal of any
computational approaches. All these drug association-related processes and
224
N. A. Murugan et al.
pragmatic to use classical mechanics to describe the molecular systems which
involves relatively simple mathematics, i.e. solving Newton’s equation of motion to
describe the interaction within system and their association with other systems and to
model their time evolution and their response to external thermodynamic variables
like temperature and pressure. As per classical mechanics, once we have the
force-field information for a system, its entire future and past can be predicted by
solving equation of motion. Force-fields can be developed by using various available
structure databases and thermodynamics data. In this chapter, We briefly cover
available force-field methods for computing the binding affinity in order to rank
protein–ligand complexes in drug discovery and design. In addition, briefly discuss
their limitations and also present the recent advancement in computational modelling
approaches based on quantum mechanical theory and machine learning algorithm in
a way suitable for drug discovery applications.
3 Free Energies Relevant to Describe Potency
and Pharmacokinetics
Free energy is the key variable that dictates the structure of biomacromolecular
complexes (protein–ligand protein, membrane–ligand, DNA–ligand etc.) and controls various molecular association and ligand transport processes. When there are
many structures possible, the one with least free energy is the most stable one.
Moreover, biomacromolecular or molecular association processes such as drug
binding to receptor, protein–protein binding and drug transport involve the minimization of Gibbs free energy (DG). Any process that involves lowering of Gibbs
free energy can proceed spontaneously. By calculating the free energy change, we
can predict whether an association process is feasible or not. In the case of a drug,
the most relevant aspect is to understand its binding affinity or potency towards a
target biomacromolecule and its association with transport proteins like albumins
[9] and metabolizing enzymes such as cytochrome P450s (CYP) [10–12] and with
glycoproteins responsible for absorption. The ligand binding to target biomacromolecule, transport protein and metabolizing enzymes is dictated by the change in
free energy of the ligand bound to these targets when compared to that in aqueous
solution. Further, it is also necessary to understand whether the compounds will
pass through certain cell membranes and also how well it will dissociate once it is
taken through oral dose which is dependent on the physicochemical properties like
lipophilicity [13] and aqueous solubility [14]. Schematic representation of various
PK computational modeling is shown in Fig. 1, and free energy of relevance is
provided in Table 1.
Computing the free energy of binding of the drug with a biological target and
other targets (such as glycoproteins, albumins, cell membranes) that mediate the
drug transport across the body to relevant target area is the main goal of any
computational approaches. All these drug association-related processes and
224
N. A. Murugan et al.