Interfacial Engineering of MoS(2) Thin Films for Wettability-Dependent Resistive Switching and Neuromorphic Behaviors.
Authors: Yang Y, Yu Y, Cui C, Liu X, Li Y, Liu P, Hui F
Journal: Nanomaterials (Basel, Switzerland)
depression treatment
mental health
open access
Abstract
Targeting
protein–protein interactions (PPIs) using small
molecules remains a formidable challenge in computational drug discovery
due to the prevalence of broad, solvent-exposed, and topographically
shallow binding interfaces. Unlike enzymes or receptors
that typically feature well-defined, deep orthosteric cavities, PPI
surfaces frequently lack deep concave pockets suitable for classical
structure-based ligand design. The costimulatory
immune receptor CD28 serves as a prime example
of such a structurally demanding PPI target. CD28 is a homodimeric
immunoglobulin-like receptor that engages the ligands CD80 and CD86
through a largely planar and solvent-accessible extracellular β-sandwich
surface. The absence of a deeply recessed cavity
in this canonical interface makes CD28 an instructive model system
for evaluating computational strategies aimed at exploiting shallow
surface depressions and partially lipophilic microenvironments. These architectural constraints have historically
limited the development of small-molecule modulators and position
CD28 as a structurally informative model for structure-guided targeting
of flat immune receptor interfaces. To productively
target such intractable interfaces with small molecules,
systematic structural interrogation is essential to identify transient
or subtle ligandable microenvironments. Concurrent with these structural analyses, recent computational
advances emphasize that screening ultralarge molecular databases significantly
increases the probability of discovering novel chemotypes for challenging
targets. However, deploying traditional
structure-based virtual screening (SBVS) workflows across billions
of compounds imposes prohibitive computational bottlenecks. To overcome
these scaling limitations, we utilized PyRMD2Dock within the PyRMD
Studio suite. This comprehensive graphical
user interface (GUI) streamlines sophisticated AI workflows, allowing
researchers to perform both ligand-based and structure-based virtual
screening without requiring advanced coding expertise. PyRMD2Dock
bridges the gap between AI-driven ligand-based virtual screening (LBVS)
and SBVS by coupling the machine learning (ML) classifier PyRMD with
the high-performance docking engine AutoDock-GPU (AD4-GPU). Furthermore, PyRMD Studio incorporates critical code optimizations
that increase screening speed by more than 3-fold compared with earlier
versions.
protein–protein interactions (PPIs) using small
molecules remains a formidable challenge in computational drug discovery
due to the prevalence of broad, solvent-exposed, and topographically
shallow binding interfaces. Unlike enzymes or receptors
that typically feature well-defined, deep orthosteric cavities, PPI
surfaces frequently lack deep concave pockets suitable for classical
structure-based ligand design. The costimulatory
immune receptor CD28 serves as a prime example
of such a structurally demanding PPI target. CD28 is a homodimeric
immunoglobulin-like receptor that engages the ligands CD80 and CD86
through a largely planar and solvent-accessible extracellular β-sandwich
surface. The absence of a deeply recessed cavity
in this canonical interface makes CD28 an instructive model system
for evaluating computational strategies aimed at exploiting shallow
surface depressions and partially lipophilic microenvironments. These architectural constraints have historically
limited the development of small-molecule modulators and position
CD28 as a structurally informative model for structure-guided targeting
of flat immune receptor interfaces. To productively
target such intractable interfaces with small molecules,
systematic structural interrogation is essential to identify transient
or subtle ligandable microenvironments. Concurrent with these structural analyses, recent computational
advances emphasize that screening ultralarge molecular databases significantly
increases the probability of discovering novel chemotypes for challenging
targets. However, deploying traditional
structure-based virtual screening (SBVS) workflows across billions
of compounds imposes prohibitive computational bottlenecks. To overcome
these scaling limitations, we utilized PyRMD2Dock within the PyRMD
Studio suite. This comprehensive graphical
user interface (GUI) streamlines sophisticated AI workflows, allowing
researchers to perform both ligand-based and structure-based virtual
screening without requiring advanced coding expertise. PyRMD2Dock
bridges the gap between AI-driven ligand-based virtual screening (LBVS)
and SBVS by coupling the machine learning (ML) classifier PyRMD with
the high-performance docking engine AutoDock-GPU (AD4-GPU). Furthermore, PyRMD Studio incorporates critical code optimizations
that increase screening speed by more than 3-fold compared with earlier
versions.