Docking Score ML: Target-Specific Machine Learning Models Improving Docking-Based Virtual Screening in 155 Targets.

In drug discovery, molecular docking methods face challenges in accurately predicting energy. Scoring functions used in molecular docking often fail to simulate complex protein-ligand interactions fully and accurately leading to biases and inaccuracies in virtual screening and target predictions. We introduce the "Docking Score ML", developed from an analysis of over 200,000 docked complexes from 155 known targets for cancer treatments. The scoring functions used are founded on bioactivity data sourced from ChEMBL and have been fine-tuned using both supervised machine learning and deep learning techniques. We validated our approach extensively using multiple data sets such as validation of selectivity mechanism, the DUDE, DUD-AD, and LIT-PCBA data sets, and performed a multitarget analysis on drugs like sunitinib. To enhance prediction accuracy, feature fusion techniques were explored. By merging the capabilities of the Graph Convolutional Network (GCN) with multiple docking functions, our results indicated a clear superiority of our methodologies over conventional approaches. These advantages demonstrate that Docking Score ML is an efficient and accurate tool for virtual screening and reverse docking.

Uncovering the selectivity mechanism of phosphodiesterase 7A/8A inhibitors through computational studies.

The expression of phosphodiesterase 7A (PDE7A) and phosphodiesterase 8A (PDE8) genes is integral to human signaling pathways, and the inhibition of PDE7A has been associated with the onset of various diseases, including effects on the immune system and nervous system. The development of PDE7 selective inhibitors can promote research on immune and nervous system diseases, such as multiple sclerosis, chronic inflammation, and autoimmune responses. PDE8A is expressed alongside PDE8B, and its inhibitory mechanism is still unclear. Studying the mechanisms of selective inhibitors against different PDE subtypes is crucial to prevent potential side effects, such as nausea and cardiac toxicity, and the sequence similarity of the two protein subtypes was 55.9%. Therefore, it is necessary to investigate the differences of both subtypes' ligand binding sites. Selective inhibitors of two proteins were chosen to summarize the reason for their selectivity through molecular docking, molecular dynamics simulation, alanine scanning mutagenesis, and MM-GBSA calculation. We found that Phe384PDE7A, Leu401PDE7A, Gln413PDE7A, Tyr419PDE7A, and Phe416PDE7A in the active site positively contribute to the selectivity towards PDE7A. Additionally, Asn729PDE8A, Phe767PDE8A, Gln778PDE8A, and Phe781PDE8A positively contribute to the selectivity towards PDE8A.

Discovery of novel CDK9 inhibitor with tridentate ligand: Design, synthesis and biological evaluation.

Cyclin-dependent kinase 9 (CDK9) plays a role in transcriptional regulation, which had become an attractive target for discovery of antitumor agent. In this work, beyond traditional CDK9 inhibitor with bidentate ligands in ATP binding domain, a series of novel CDK9 inhibitor with tridentate ligand were designed and synthesized. Surprisingly, this unique tridentate ligand structure endows better CDK9 inhibition selectivity compared to other CDK subtypes, and the lead candidate compound Z4-7a showed effective proliferation inhibition in HCT116 cells with acceptable pharmacokinetic properties. Research on the mechanism indicated that Z4-7a could induce apoptosis in the HCT116 cell line by inhibiting phosphorylation of RNA polymerase II at Ser2, which resulted in the inhibition of apoptosis-related genes and proteins expression. In brief, introduction of tridentate ligand might work as a promising strategy for the development of novel selective CDK9 inhibitor.

Selectivity mechanism of BCL-XL/2 inhibition through in silico investigation.

The BCL-XL protein is among the most important members of the anti-apoptotic subfamily of the BCL-2 protein family, and is currently a promising new target for anti-tumor drug research. However, the BCL-XL/2 proteins have similar structures and functions, which could lead to undesirable side effects because of inhibitors that can bind to both BCL-XL and BCL-2. Therefore, it is crucial to expound on the structural basis of the selective mechanism towards BCL-XL/2 inhibition. In the current study, we employed hybrid computational methods including molecular docking and dynamics simulation, MM/GBSA energy calculation, alanine scanning mutagenesis and Hirshfeld surface analysis to comprehensively reveal the selectivity mechanism towards BCL-XL/2 from multiple perspectives, revealing the significant effects of the BCL-XL residues SER106 and LEU108 as well as the BCL-2 residue ASP103 on the inhibitory selectivity. Overall, our findings provide useful references for the rational design of BCL-XL/2 selective inhibitors with better affinity.

Selectivity mechanism of muscarinic acetylcholine receptor antagonism through in silico investigation.

Structures of muscarinic acetylcholine receptors illustrate the strikingly high degree of homology of the residues among isoforms, thus leading to difficulty in achieving subtype selectivity when targeting these receptors and causing undesired side effects when treating the corresponding diseases. Considering the urgent need for more selective and potency therapies, this study is aimed at revealing the selectivity mechanism against M4/5 via in silico strategies, revealing crucial molecular interactions such as hydrogen bonds and pi-cation interaction formed between the key residues TYR416, ASN417, and TRP435 of M4, respectively, hydrophobic pocket formed by the key residues, especially CYS484 of M5. Besides, the water around TYR416M4 and ASN459M5, which can be replaced by substituent groups which can form the hydrogen bond interaction network by simulated bridging water and the water around ASP112M4, whose replacement maybe not contribute to the increase in binding affinity of the compound, may affect the inhibitory selectivity among M4/5 in the aspect of the solvent. Moreover, from the point of inhibitors, compounds with a positively ionizable group could selectively bind to M4 receptors, while hydrophobic molecules may bind preferably to M5. We believe that the current study would provide a basis for the design of subsequent M4/5 selective antagonists.

Perspective of structural flexibility on selective inhibition towards CYP1B1 over CYP1A1 by α-naphthoflavone analogs.

Research on action selectivity between CYP1A1 and CYP1B1 is particularly valuable for cancer chemoprevention and chemotherapy. However, they share a very close similarity in their ligand-binding pockets that α-naphthoflavone (ANF) is the co-crystal ligand for both isoforms, which poses a major challenge in revealing their selectivity mechanism. Therefore, three selective CYP1B1 inhibitors derived from ANF were selected to illustrate the structural basis for the selectivity between the two isoforms via a comprehensive computational strategy. It was found that the sustainability of the π-π stacking interactions with the phenylalanine residues of the two isoforms, namely, Phe123, Phe224, and Phe258 for CYP1A1, and Phe134, Phe231, and Phe268 for CYP1B1, played a crucial role in determining the selectivity of ligands with a classic aromatic conjugation system like ANF and its derivatives for CYP1B1 versus CYP1A1. Of note, the structural flexibility of the corresponding protein domains mainly orchestrated the sustainability of the corresponding π-π stacking interactions, thereby determining the binding selectivity. Therefore, the structure modification of naphthoflavone lead compounds into preferable binding configurations to satisfy the π-π stacking interactions of the key phenylalanine residues within CYP1B1 would be an inspiring strategy devised to improve the inhibitory selectivity towards CYP1B1. Collectively, this study revealed valuable insight into understanding the selective mechanism between CYP1A1 and CYP1B1 from the perspective of structural flexibility, which sheds light on the future rational design of CYP1B1 selective inhibitors.

Rational design of selective HDAC2 inhibitors for liver cancer treatment: computational insights into the selectivity mechanism through molecular dynamics simulations and QM/MM calculations.

The rational design of selective histone deacetylase 2 (HDAC2) inhibitors is beneficial for the therapeutic treatment of liver cancer, though HDAC2 is highly homologous to HDAC8, which may lead to undesired side effects due to the pan-inhibition towards HDAC2 and HDAC8. To clarify the structural basis of selective inhibition towards HDAC2 over HDAC8, we utilized multiple in silico strategies, including sequence alignment, structural comparison, molecular docking, molecular dynamics simulations, free energy calculations, alanine scanning mutagenesis, pharmacophore modeling, protein contacts atlas analysis and QM/MM calculations to study the binding patterns of HDAC2/8 selective inhibitors. Through the whole process described above, it is found that although HDAC2 has conserved GLY154 and PHE210 that also exist within HDAC8, namely GLY151 and PHE208, the two isoforms exhibit diverse binding modes towards their inhibitors. Typically, HDAC2 inhibitors interact with the Zn2+ ions through the core chelate group, while HDAC8 inhibitors adopt a bent conformation within the HDAC8 pocket that inclines to be in contact with the Zn2+ ions through the terminal hydroxamic acid group. In summary, our data comprehensively elucidate the selectivity mechanism towards HDAC2 over HDAC8, which would guide the rational design of selective HDAC2 inhibitors for liver cancer treatment.

Identification of Pyruvate Dehydrogenase E1 as a Potential Target against Magnaporthe oryzae through Experimental and Theoretical Investigation.

Magnaporthe oryzae (M. oryzae) is a typical cause of rice blast in agricultural production. Isobavachalcone (IBC), an active ingredient of Psoralea corylifolia L. extract, is an effective fungicide against rice blast. To determine the mechanism of IBC against M. oryzae, the effect of IBC on the metabolic pathway of M. oryzae was explored by transcriptome profiling. In M. oryzae, the expression of pyruvate dehydrogenase E1 (PDHE1), part of the tricarboxylic acid (TCA cycle), was significantly decreased in response to treatment with IBC, which was verified by qPCR and testing of enzyme activity. To further elucidate the interactions between IBC and PDHE1, the 3D structure model of the PDHE1 from M. oryzae was established based on homology modeling. The model was utilized to analyze the molecular interactions through molecular docking and molecular dynamics simulation, revealing that IBC has π-π stacking interactions with residue TYR139 and undergoes hydrogen bonding with residue ASP217 of PDHE1. Additionally, the nonpolar residues PHE111, MET174, ILE 187, VAL188, and MET250 form strong hydrophobic interactions with IBC. The above results reveal that PDHE1 is a potential target for antifungal agents, which will be of great significance for guiding the design of new fungicides. This research clarified the mechanism of IBC against M. oryzae at the molecular level, which will underpin further studies of the inhibitory mechanism of flavonoids and the discovery of new targets. It also provides theoretical guidance for the field application of IBC.

Computational insights into the sorption mechanism of environmental contaminants by carbon nanoparticles through molecular dynamics simulation and density functional theory.

Carbon nanomaterials like carbon nanotubes (CNTs), graphene or graphene oxide (GO) could significantly enhance contaminant sorption in aqueous solutions, and offer a promising opportunity in water and air purification for the removal of environmental contaminants. Although the sorption of environmental contaminants by carbon nanoparticles has been reported in many studies, the knowledge regarding their molecular mechanism at the atomic level is still very limited. In this study, we integrate density functional theory (DFT) calculation, fully atomistic molecular dynamics (MD) simulation and binding energy calculation to investigate the sorption mechanism of environmental contaminants by carbon nanoparticles. We proposed that CNTs and graphene-carbon nanoparticles can be used for the sorption of hydrophobic contaminants, and the GO-carbon nanoparticles are more suitable for the sorption of polar and ionic contaminants, and we believe that the results would have important implications for the future application of carbon nanoparticles in environmental pollution cleanup.

A novel cathepsin L inhibitor prevents the progression of idiopathic pulmonary fibrosis.

In previous work, the target of asperphenamate as a natural product was successfully determined as cathepsin by the natural product consensus pharmacophore strategy. In order to develop accurate SAR (structure-activity relationship) of asperphenamate-type cathepsin inhibitor, we chose several novel analogs with heterocyclic moiety to perform further study. The molecular simulation showed that 4-pyridyl derivative 3 with the greatest cathepsin inhibitory activity presented new interaction modes with protein utilizing its B-ring moiety. And then molecular dynamics (MD) simulation further revealed that 3 and cathepsin kept stable interaction in the binding site, which validated the molecular docking results. In view that cathepsins play an important role in fibrosis and some cathepsin inhibitors display the therapeutic ability for fibrosis, we investigated the anti-fibrotic effect of 3in vitro and in vivo. The results indicated that 3 displayed the strongest inhibitory effect on the formation of α-SMA and collagen I as the protein markers of fibrosis among all tested derivatives. Further in vivo assay confirmed that 3 indeed showed significant inhibitory ability against pulmonary fibrosis by the method of H&E and Masson staining as well as immunohistochemical staining for characteristic α-SMA proteins.

Preparation and application of novel MIL-101(Cr) composite in liquid chromatographic separation of aromatic compounds: experimental and computational insights.

A novel composite material of SiO2@dSiO2@MIL-101(Cr) was synthesized via SiO2@dSiO2 as the core and MIL-101(Cr) as the shell to separate aromatic compounds. The laboratory prepared column gave rise to the baseline separation of xylene, dichlorobenzene isomers, phthalate esters, nitrobenzene, and acetophenone with high column efficiency (e.g., 109,050 plates m-1 for methyl phthalate) and good precision (e.g., 0.02-0.05%, 0.24-0.34%, 0.14-0.18%, and 0.11-0.13% corresponding to the relative standard deviation of retention time, peak area, peak height, and half peak width for xylene isomers, respectively). The calculation of thermodynamic parameters demonstrated that the separation of o-xylene, nitrobenzene, acetophenone, and p-dichlorobenzene was controlled by positive ∆H and ∆S. Although the separation of aromatic compounds by a MOF packed column has been reported in many studies, the knowledge regarding their separation mechanism at atomic level is still very limited. In this study, we integrate fully atomistic molecular dynamics simulation and binding free energy calculation to investigate the separation mechanism of aromatic compounds by MIL-101(Cr). The investigation provides a base for separation of more and other compounds in the future. Graphical abstract.

In Silico Exploration of the Molecular Mechanism of Cassane Diterpenoids on Anti-inflammatory and Immunomodulatory Activity.

Cassane diterpenoids (CAs), recognized as main constituents of many medical plants of the genus Caesalpinia, exhibit diverse bioactivities, including anti-inflammatory and immunomodulatory activity, and also showed a therapeutic effect on rheumatoid arthritis (RA) according to previous work, including ours. In this study, 102 CA compounds were selected to explore the possible molecular mechanism of this class of natural products on anti-inflammatory and immunomodulatory activity using RA as a disease model through a series of in silico methods: chemical-similarity-based target prediction, molecular docking, and molecular dynamics (MD) simulation. As a consequence, four signaling pathways (TCR signaling pathway, TLR signaling pathway, VEGF signaling pathway, and osteoclast differentiation pathway) by which CAs exert their effect on inflammation and immunomodulation were identified. Furthermore, the binding modes of CAs complexing with several crucial targets, which were picked out by credible docking results and took part in these signaling pathways, were explored by MD simulations. This is the first time that the molecular mechanism of the anti-RA activity of natural CAs has been investigated with in silico methods, and these findings might explain the activity of CAs on anti-inflammation and immunomodulation, which could supply a valuable reference for drug design research on CAs.

Molecular dynamics simulation and QM/MM calculation reveal the selectivity mechanism of type I 1/2 kinase inhibitors: the effect of intramolecular H-bonds and conformational restriction for improved selectivity.

Understanding the selectivity mechanisms of inhibitors towards highly similar proteins is extremely important work on the way to a new drug. Here, we aim to reveal the selectivity mechanisms of type I 1/2 kinase inhibitors towards p21-activated kinase (PAK4) and mitogen-activated protein kinase kinase kinase 14 (MAP3K14, NIK). PAK4, belonging to the serine/threonine protein kinases, is involved in cell signaling pathways and controls cellular functions and has received attention as an attractive drug target. The high sequence identity between PAK4 and NIK makes it challenging to design selective PAK4 inhibitors. In this work, computational methods including protein comparison, molecular docking, QM/MM, molecular dynamics simulations, and density functional theory (DFT) calculation were employed to explore the binding mechanisms of selective inhibitors against NIK and PAK4. The simulation results revealed the crucial factors accounting for selective inhibition of PAK4 over NIK, including different protein-ligand interactions, the positions and conformations of key residues, and the ligands flexibilities. This study will shed light on understanding the selectivity mechanisms of PAK4 and NIK inhibitors.

Discovery of a novel cathepsin inhibitor with dual autophagy-inducing and metastasis-inhibiting effects on breast cancer cells.

Drug resistance and cancer cells metastasis have been the leading causes of chemotherapy failure and cancer-associated death in breast cancer patients. In present, various active molecules either exhibiting novel mechanism of action such as inducing autophagy or inhibiting metastasis have been developed to address these problems. However, the compounds exhibiting such dual functions have rarely been described. Previous work in our group showed that TSA, as a synthetic analog of asperphenamate, induced autophagic cell death in breast cancer cells instead of apoptosis. Furthermore, the target enzyme of TSA was predicted to be cathepin L (Cat L) by natural product consensus pharmacophore strategy. Accumulated evidences have shown that cathepsins are closely associated with migration and invasion of breast cancer cells. It seemed likely that TSA-like molecules may possess the dual functions of inducing autophagy and inhibiting metastasis. Therefore, sixty optically active derivatives were firstly designed and synthesized by replacing the A-ring moiety of TSA with other substituted-phenyl sulfonyl groups. Further cathepsin inhibitory activity assay showed that (S, S) and (S, R) isomers displayed no activity against four kinds of cathepsins (L, S, K, B), while all derivatives tested were inactive toward K and B subtypes. Compound 6a with meta-bromo substituent displayed the greatest inhibitory activity, and its inhibitory capability against Cat L and S was 3.9 and 11.5-fold more potent than that of TSA, respectively. Molecular docking also exhibited that 6a formed more hydrogen bonds or π-π contacts with Cat L or S than TSA. In order to determine whether 6a could play dual roles, its anti-cancer mechanism was further investigated. On the one hand, MDC staining experiment and western blotting analysis validated that 6a can induce autophagy in MDA-MB-231 cells. On the other hand, its metastatic inhibitory ability was also confirmed by wound healing and transwell chamber experiment.

Pyrrolo[2,1-c][1,4] benzodiazepine-3,11-diones protect SHSY-5Y cells from Cd-induced apoptosis involving suppression of endoplasmic reticulum stress.

Cadmium (Cd) is a potent toxic heavy metal, some studies showed that Cd-induced apoptosis is through ER stress pathway. Compounds of pyrrolo[2,1-c][1,4]benzodiazepine (PBD)-3,11-diones were discovered as potent neuroprotective agents against Cd-induced toxicity in SH-SY5Y cells for the first time. In this study, twenty-six PBD-3,11-dione derivatives were synthesized and evaluated for their neuroprotective activity against Cd-induced toxicity by CCK-8 assay. Their preliminary SARs studies indicated that various substituents were tolerated on the benzene ring, and alkyl heterocycles groups at the N10-position of the PBD-3,11-dione scaffold were important for the activities. Among them, compound 13c exhibited the best activity (cell viability = 68.6%, 25 µM). Furthermore, we found that the compound 13c could inhibit cadmium-induced cell apoptosis with the downregulation of the ER stress markers GRP78, CHOP, cleaved-caspase12 and cleaved-caspase3 through western blotting. The results of in silico evaluation of ADME/T properties showed that 13c exhibited medium BBB penetration level and promising toxicity profiles. These results proved the potential of 13c as a promising lead compound against Cd-induced neurotoxicity.

Discovery of novel cathepsin inhibitors with potent anti-metastatic effects in breast cancer cells.

It is still challenging to determine the potential targets of natural products, which is essential for further drug research and development. Due to its novel mechanism of action of inducing autophagy effects in breast cancer cells, asperphenamate has received our considerable attention. However, its unknown target inevitably impedes further study. In our previous work, the target enzyme of asperphenamate was predicted as cathepsin by the natural product consensus pharmacophore strategy. Then, asperphenamate and its three derivatives were chosen to study in detail by molecular docking calculations with AutoDock 4 suite. The docking results showed the three derivatives interacted more tightly with either cathepsin L or cathepsin S than with asperphenamate. The ortho-benzyloxyl phenylacetyl derivative 1 andp-toluenesulfonyl derivative 3 showed similar interactions with cathepsin L and adopted a better geometric shape within the binding pocket than did the N-CBZ-piperidyl analog 2. On the other hand, compound 2 formed more hydrogen bonds than 1 and 3 to make it tightly bind within cathepsin S. The cathepsin inhibitory activity assay verified the molecular simulation results. Compound 2 was remarkably less active than 1 and 3 against cathepsin L. However, compound 2 showed the strongest potency against cathepsin S with IC50 of 13.12 ±â€¯0.29 µM. Considering that cathepsin S plays a vital role in the process of metastasis in breast cancer cells, the inhibitory effect of 2 on migration and invasion was further studied in human breast cancer MDA-MB-231 cells by wound healing and transwell chamber assays. The results illustrated that 2 exhibited an apparent inhibitory ability to the metastasis of MDA-MB-231 cells. Next, 2 will be chosen as a lead compound to develop novel double functional chemotherapeutic agents with both novel mechanisms of action against apoptosis-resistant cancer cells, such as inducing autophagy and inhibiting cancer metastasis.

An amino-functionalized magnetic framework composite of type Fe3O4-NH2@MIL-101(Cr) for extraction of pyrethroids coupled with GC-ECD.

An amino-functionalized magnetic framework composite of type Fe3O4-NH2@MIL-101(Cr) was synthesized using a solvothermal method. The material was characterized by scanning electron microscopy, X-ray diffraction, Fourier transform infrared spectroscopy, nitrogen adsorption, and magnetometry. The composite combines the advantages of amino-modified Fe3O4 and MIL-101(Cr). The presence of amino groups facilitates the fairly specific adsorption of pyrethroids. The composite was employed as a sorbent for magnetic solid phase extraction of five pyrethroids from environmental water samples. Following desorption with acidified acetone, the pyrethroids were quantified by gas chromatography with electron capture detection. The detection limits for bifenthrin, fenpropathrin, λ-cyhalothrin, permethrin, and deltamethrin range from 5 to 9 pg·mL-1. The method is rapid, accurate, and highly sensitive. The molecular interactions and free binding energies between MIL-101(Cr) and the five pyrethroids were calculated by means of molecular docking. Graphical abstract A novel functionalized magnetic framework composite of type Fe3O4-NH2@MIL-101(Cr) was synthesized. It was applied as a sorbent for magnetic solid phase extraction of pyrethroids prior to their quantitation by gas chromatography with electron capture detection. The molecular interactions of analytes and MIL-101(Cr) were studied.

Structural simulation and selective inhibitor discovery study for histone demethylases KDM4E/6B from a computational perspective.

The methylation and demethylation of lysine and arginine side chains are fundamental processes in gene regulation and disease development. Histone lysine methylation, controlled by histone lysine methyltransferases (KMTs) and histone lysine demethylases (KDMs), plays a vital role in maintaining cellular homeostasis and has been implicated in diseases such as cancer and aging. This study focuses on two members of the lysine demethylase (KDM) family, KDM4E and KDM6B, which are significant in gene regulation and disease pathogenesis. KDM4E demonstrates selectivity for gene regulation, particularly concerning cancer, while KDM6B is implicated in inflammation and cancer. The study utilizes specific inhibitors, DA-24905 and GSK-J1, showcasing their exceptional selectivity for KDM4E and KDM6B, respectively. Employing an array of computational simulations, including sequence alignment, molecular docking, dynamics simulations, and free energy calculations, we conclude that although the binding cavities of KDM4E and KDM6B has high similarity, there are still some different crucial amino acid residues, indicating diverse binding forms between protein and ligands. Various interaction predominates when proteins are bound to different ligands, which also has significant effect on selective inhibition. These findings provide insights into potential therapeutic strategies for diseases by selectively targeting these KDM members.

FaissMolLib: An efficient and easy deployable tool for ligand-based virtual screening.

Virtual screening-based molecular similarity and fingerprint are crucial in drug design, target prediction, and ADMET prediction, aiding in identifying potential hits and optimizing lead compounds. However, challenges such as lack of comprehensive open-source molecular fingerprint databases and efficient search methods for virtual screening are prevalent. To address these issues, we introduce FaissMolLib, an open-source virtual screening tool that integrates 2.8 million compounds from ChEMBL and ZINC databases. Notably, FaissMolLib employs the highly efficient Faiss search algorithm, outperforming the Tanimoto algorithm in identifying similar molecules with its tighter clustering in scatter plots and lower mean, standard deviation, and variance in key molecular properties. This feature enables FaissMolLib to screen 2.8 million compounds in just 0.05 seconds, offering researchers an efficient, easily deployable solution for virtual screening on laptops and building unique compound databases. This significant advancement holds great potential for accelerating drug discovery efforts and enhancing chemical data analysis. FaissMolLib is freely available at http://liuhaihan.gnway.cc:80. The code and dataset of FaissMolLib are freely available at https://github.com/Superhaihan/FiassMolLib.

Insights into the selective mechanism of PDE2/9a inhibitors from silico aspects.

The selective design of competitive enzyme inhibitors is an extremely difficult task but necessary work for certain types of systems, such as the phosphodiesterase (PDE) system addressed in this article. In the PDE family, PDE2A and PDE9 respectively target the central nervous system and heart failure, and share many conserved amino acids at their binding sites. Therefore, gaining a deep understanding of the selective mechanisms of PDE2A/9A is crucial for designing highly selective drugs. In this study, various computer-aided drug design (CADD) methods, including molecular docking, molecular dynamics simulations (MD), and binding free energy calculations, are employed to explore the selective mechanisms of PDE2A/9A. Overall, our research results indicate a selective design strategy for PDE2A, which involves incorporating hydrophobic or aromatic moieties into the molecular structure to better accommodate the hydrophobic pocket of PDE2A. Additionally, it is recommended to introduce functional groups capable of forming connections with selective residues, such as Phe830 and Gln812 for PDE2A, or Ala452 and Tyr424 for PDE9A, to enhance the selectivity of inhibitors targeting PDE2A/9A. This achievement is anticipated to pave the way for the development of innovative and selective small molecules targeting PDE2A/9A.Communicated by Ramaswamy H. Sarma.