time, further involvement of research groups with varied backgrounds and availability of clean data are expected to inspire emergence of more efficient workflows in
drug design that combine traditional methods and machine learning methods.
8 Conclusions
The current drug discovery projects can be benefited a lot from advancements in the
structure elucidation methods such as cryogenic electron microscopy, NMR
spectroscopy, X-ray crystallograpy and from computational free energy calculation
methods. This chapter presents various computational approaches available for
estimating the free energies of drugs in different environments which surrogate
various components of fate of drugs in the biosystem and this not only limited to drug
binding to its targets but also other interactions and its relevant properties e.g.
ADMET. We present various free energy calculation methods which use force-field,
semi-empirical and ab initio electronic structure theory-based methods. Until a
decade ago, using the electronic structure theory method for studying the structure
and energetics of biomacromolecule was formidable. Thanks to the fragmentation
and effective Hamiltonian approaches, it is possible to employ these methods for
computing the interaction energy between ligand and biomacromolecular fragments
reliably. Here, various working principles of these approaches along with key
illustrative examples are presented. Even though the methods appear very promising
for computing the free energy of the ligands in solvent or in biomacromolecular (such
as enzyme, membrane, fibril, DNA and RNA) environment, we need to systematically study various receptor–ligand systems and test for their ability to reproduce
experimental binding affinity and other pharmacokinetic parameters before
employing them as lead compounds in drug discovery projects. While physics-based
methods such as those mentioned above are important and unavoidable, alternative
approaches based on machine learning algorithms that exploit existing experimental/
computational data are emerging to be powerful tools for drug design. We expect that
elegant combination of traditional physics-based methods, better computational
power and more sophisticated machine learning algorithms will enable efficient and
accurate quantification of protein–ligand binding affinities for improved lead
identification/optimization processes in the drug design and discovery projects.
References
1. Rask-Andersen M, Almen MS, Schioth HB (2011) Trends in the exploitation of novel drug
targets. Nat Rev Drug Discov 10:579–590
2. Lenz GR, Nash HM, Jindal S (2000) Chemical ligands, genomics and drug discovery. Drug
Discov Today 5(4):145–156
3. Knowles J, Gromo G (2003) Target selection in drug discovery. Nat Rev Drug Discov 2:
63–69
Recent Advancements in Computing Reliable Binding Free Energies …
243
drug design that combine traditional methods and machine learning methods.
8 Conclusions
The current drug discovery projects can be benefited a lot from advancements in the
structure elucidation methods such as cryogenic electron microscopy, NMR
spectroscopy, X-ray crystallograpy and from computational free energy calculation
methods. This chapter presents various computational approaches available for
estimating the free energies of drugs in different environments which surrogate
various components of fate of drugs in the biosystem and this not only limited to drug
binding to its targets but also other interactions and its relevant properties e.g.
ADMET. We present various free energy calculation methods which use force-field,
semi-empirical and ab initio electronic structure theory-based methods. Until a
decade ago, using the electronic structure theory method for studying the structure
and energetics of biomacromolecule was formidable. Thanks to the fragmentation
and effective Hamiltonian approaches, it is possible to employ these methods for
computing the interaction energy between ligand and biomacromolecular fragments
reliably. Here, various working principles of these approaches along with key
illustrative examples are presented. Even though the methods appear very promising
for computing the free energy of the ligands in solvent or in biomacromolecular (such
as enzyme, membrane, fibril, DNA and RNA) environment, we need to systematically study various receptor–ligand systems and test for their ability to reproduce
experimental binding affinity and other pharmacokinetic parameters before
employing them as lead compounds in drug discovery projects. While physics-based
methods such as those mentioned above are important and unavoidable, alternative
approaches based on machine learning algorithms that exploit existing experimental/
computational data are emerging to be powerful tools for drug design. We expect that
elegant combination of traditional physics-based methods, better computational
power and more sophisticated machine learning algorithms will enable efficient and
accurate quantification of protein–ligand binding affinities for improved lead
identification/optimization processes in the drug design and discovery projects.
References
1. Rask-Andersen M, Almen MS, Schioth HB (2011) Trends in the exploitation of novel drug
targets. Nat Rev Drug Discov 10:579–590
2. Lenz GR, Nash HM, Jindal S (2000) Chemical ligands, genomics and drug discovery. Drug
Discov Today 5(4):145–156
3. Knowles J, Gromo G (2003) Target selection in drug discovery. Nat Rev Drug Discov 2:
63–69
Recent Advancements in Computing Reliable Binding Free Energies …
243