bridge the gap to longer timescales. Another effective combination
is that of SCA analysis and MD [121, 163]. Platforms for analyses
of structures and trajectories have started to emerge
[164, 165]. These combined platforms allow for rapid pipelining
of combinations of advanced methods to achieve high quality
models.5 Conclusions
Understanding how protein drug targets function is very complex
indeed. Many experienced drug hunters opine that there is much
serendipity involved in discovering a drug candidate. At the level of
drug target molecular mechanisms, the continual progress in
molecular dynamics, in terms of timescales and model accuracy,
offers a new level of rationality. In the future complex molecular
mechanisms might be identified and understood at earlier stages in
the drug discovery process, allowing valuable new functional effects
and fine-tuned selectivity.
It is clear that atomistic, explicitly solvated molecular dynamics
models offer a good balance between sampling speed and transfer-
able accuracy. However, detailed aspects of protein drug targets
such as protonation states and metal ion binding are especially
important for delicate dynamic transitions, and in some cases
require advanced and computationally more expensive models.
Also, building models from crystal structures can be challenging
even for experienced practitioners. Details about the local environ-
ment of the protein such as oxidation, lipids, and glycosylation are
often crucial. In many cases mistakes made in initial model building
can cause artifactual or misleading simulation data.
Despite the well-known limitations of molecular dynamics
models, finite simulation length and underlying model quality,
there are now many examples where simulations offer unique
understanding of important dynamic transitions in protein drug
targets. Crucially, these models are now often validated and some-
times driven by experimental data. Highlights discussed here are for
important kinase, GPCR, and protease targets.
Many pharmaceutical companies now have specialists in molec-
ular dynamics simulations within their research departments along
with some significant high-performance computing capabilities.
However, the majority of these researchers focus on small molecule
binding affinity calculations such as those developed by Schrodin-
ger. DESRES is a well-known leader in applying molecular dynam-
ics to conformational transitions of drug targets. Another small
company called Accellera is selling a specialist MSM and adaptive
sampling platform to the industry along with support and services.Computational Study of Protein Conformational Transitions 355