Houman Kazemzadeh
Houman Kazemzadeh has a total of 49 co-authors. Here are the top co-authors by number of shared publications:
Afshin Cheraghzade
Department of Occupational Health Engineering, School of Health
3 shared publicationsAli Baheri
3 shared publicationsAli Behrad
Department of Electrical Engineering
3 shared publicationsAli Reza Keivanimehr
Department of Management, Science and Technology
3 shared publicationsAli Sabzi
Department of Computer Engineering
3 shared publicationsArshia Abolghasemi
School of Electrical and Computer Engineering
3 shared publications
Latest Publications
- Open Access
- 17 September 2026
- Published by American Chemical Society (ACS) in ChemRxiv
From Computational Chemistry to Generative Models: A Survey of AI-Driven Small-Molecule Drug Discovery
Small-molecule drug discovery requires navigating enormous chemical space to identify candidates that are simultaneously potent, selective, synthetically accessible, and acceptable across pharmacokinetic and safety profiles; modern generative AI extends a long computational medicinal chemistry lineage that began with QSAR, pharmacophore modeling, and early structure-based de novo design. This survey focuses primarily on developments from 2017 to 2025, with the literature updated through 31 January 2026. Using a medicinal-chemistry lens, it first establishes the shared machinery of the field: molecular representations such as SMILES, SELFIES, graphs, fingerprints, three-dimensional coordinates, voxels, and latent vectors; commonly used datasets and benchmarks; and graph neural network encoders, including message-passing, directional, and equivariant architectures. We compare five principal generative mechanisms—autoregressive models, variational autoencoders, generative adversarial networks, normalizing flows, and diffusion models—together with reinforcement learning as a major goal-directed optimization framework. Across these approaches, we show that current methods offer complementary rather than interchangeable trade-offs, and we highlight recurring challenges, including validity versus synthesizability, limited prospective experimental validation, mode collapse and diversity loss, stereochemical and physical realism in three-dimensional design, computational cost, data scarcity, and the emerging convergence toward scaffold-conditioned generation, hybrid architectures, developability-aware optimization, and molecular foundation models.
- 17 September 2026
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessFrom Computational Chemistry to Generative Models: A Survey of AI-Driven Small-Molecule Drug Discovery
Small-molecule drug discovery requires navigating enormous chemical space to identify candidates that are simultaneously potent, selective, synthetically accessible, and acceptable across pharmacokinetic and safety profiles; modern generative AI extends a long computational medicinal chemistry lineage that began with QSAR, pharmacophore modeling, and early structure-based de novo design. This survey focuses primarily on developments from 2017 to 2025, with the literature updated through 31 January 2026. Using a medicinal-chemistry lens, it first establishes the shared machinery of the field: molecular representations such as SMILES, SELFIES, graphs, fingerprints, three-dimensional coordinates, voxels, and latent vectors; commonly used datasets and benchmarks; and graph neural network encoders, including message-passing, directional, and equivariant architectures. We compare five principal generative mechanisms—autoregressive models, variational autoencoders, generative adversarial networks, normalizing flows, and diffusion models—together with reinforcement learning as a major goal-directed optimization framework. Across these approaches, we show that current methods offer complementary rather than interchangeable trade-offs, and we highlight recurring challenges, including validity versus synthesizability, limited prospective experimental validation, mode collapse and diversity loss, stereochemical and physical realism in three-dimensional design, computational cost, data scarcity, and the emerging convergence toward scaffold-conditioned generation, hybrid architectures, developability-aware optimization, and molecular foundation models.
- Open Access
- 15 September 2026
- Published by American Chemical Society (ACS) in ChemRxiv
From Computational Chemistry to Generative Models: A Survey of AI-Driven Small-Molecule Drug Discovery
Small-molecule drug discovery requires navigating enormous chemical space to identify candidates that are simultaneously potent, selective, synthetically accessible, and acceptable across pharmacokinetic and safety profiles; modern generative AI extends a long computational medicinal chemistry lineage that began with QSAR, pharmacophore modeling, and early structure-based de novo design. This survey focuses primarily on developments from 2017 to 2025, with the literature updated through 31 January 2026. Using a medicinal-chemistry lens, it first establishes the shared machinery of the field: molecular representations such as SMILES, SELFIES, graphs, fingerprints, three-dimensional coordinates, voxels, and latent vectors; commonly used datasets and benchmarks; and graph neural network encoders, including message-passing, directional, and equivariant architectures. We compare five principal generative mechanisms—autoregressive models, variational autoencoders, generative adversarial networks, normalizing flows, and diffusion models—together with reinforcement learning as a major goal-directed optimization framework. Across these approaches, we show that current methods offer complementary rather than interchangeable trade-offs, and we highlight recurring challenges, including validity versus synthesizability, limited prospective experimental validation, mode collapse and diversity loss, stereochemical and physical realism in three-dimensional design, computational cost, data scarcity, and the emerging convergence toward scaffold-conditioned generation, hybrid architectures, developability-aware optimization, and molecular foundation models.
- 15 September 2026
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessFrom Computational Chemistry to Generative Models: A Survey of AI-Driven Small-Molecule Drug Discovery
Small-molecule drug discovery requires navigating enormous chemical space to identify candidates that are simultaneously potent, selective, synthetically accessible, and acceptable across pharmacokinetic and safety profiles; modern generative AI extends a long computational medicinal chemistry lineage that began with QSAR, pharmacophore modeling, and early structure-based de novo design. This survey focuses primarily on developments from 2017 to 2025, with the literature updated through 31 January 2026. Using a medicinal-chemistry lens, it first establishes the shared machinery of the field: molecular representations such as SMILES, SELFIES, graphs, fingerprints, three-dimensional coordinates, voxels, and latent vectors; commonly used datasets and benchmarks; and graph neural network encoders, including message-passing, directional, and equivariant architectures. We compare five principal generative mechanisms—autoregressive models, variational autoencoders, generative adversarial networks, normalizing flows, and diffusion models—together with reinforcement learning as a major goal-directed optimization framework. Across these approaches, we show that current methods offer complementary rather than interchangeable trade-offs, and we highlight recurring challenges, including validity versus synthesizability, limited prospective experimental validation, mode collapse and diversity loss, stereochemical and physical realism in three-dimensional design, computational cost, data scarcity, and the emerging convergence toward scaffold-conditioned generation, hybrid architectures, developability-aware optimization, and molecular foundation models.
- Open Access
- 1 September 2026
- Published by American Chemical Society (ACS) in ChemRxiv
From Computational Chemistry to Generative Models: A Survey of AI-Driven Small-Molecule Drug Discovery
Small-molecule drug discovery requires navigating enormous chemical space to identify candidates that are simultaneously potent, selective, synthetically accessible, and acceptable across pharmacokinetic and safety profiles; modern generative AI extends a long computational medicinal chemistry lineage that began with QSAR, pharmacophore modeling, and early structure-based de novo design. This survey focuses primarily on developments from 2017 to 2025, with the literature updated through 31 January 2026. Using a medicinal-chemistry lens, it first establishes the shared machinery of the field: molecular representations such as SMILES, SELFIES, graphs, fingerprints, three-dimensional coordinates, voxels, and latent vectors; commonly used datasets and benchmarks; and graph neural network encoders, including message-passing, directional, and equivariant architectures. We compare five principal generative mechanisms—autoregressive models, variational autoencoders, generative adversarial networks, normalizing flows, and diffusion models—together with reinforcement learning as a major goal-directed optimization framework. Across these approaches, we show that current methods offer complementary rather than interchangeable trade-offs, and we highlight recurring challenges, including validity versus synthesizability, limited prospective experimental validation, mode collapse and diversity loss, stereochemical and physical realism in three-dimensional design, computational cost, data scarcity, and the emerging convergence toward scaffold-conditioned generation, hybrid architectures, developability-aware optimization, and molecular foundation models.
- 1 September 2026
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessFrom Computational Chemistry to Generative Models: A Survey of AI-Driven Small-Molecule Drug Discovery
Small-molecule drug discovery requires navigating enormous chemical space to identify candidates that are simultaneously potent, selective, synthetically accessible, and acceptable across pharmacokinetic and safety profiles; modern generative AI extends a long computational medicinal chemistry lineage that began with QSAR, pharmacophore modeling, and early structure-based de novo design. This survey focuses primarily on developments from 2017 to 2025, with the literature updated through 31 January 2026. Using a medicinal-chemistry lens, it first establishes the shared machinery of the field: molecular representations such as SMILES, SELFIES, graphs, fingerprints, three-dimensional coordinates, voxels, and latent vectors; commonly used datasets and benchmarks; and graph neural network encoders, including message-passing, directional, and equivariant architectures. We compare five principal generative mechanisms—autoregressive models, variational autoencoders, generative adversarial networks, normalizing flows, and diffusion models—together with reinforcement learning as a major goal-directed optimization framework. Across these approaches, we show that current methods offer complementary rather than interchangeable trade-offs, and we highlight recurring challenges, including validity versus synthesizability, limited prospective experimental validation, mode collapse and diversity loss, stereochemical and physical realism in three-dimensional design, computational cost, data scarcity, and the emerging convergence toward scaffold-conditioned generation, hybrid architectures, developability-aware optimization, and molecular foundation models.
- Open Access
- 15 September 2025
- Published by Korean Pediatric Society in Clinical and Experimental Pediatrics
- Vol. 68 (9) , 700-711
Linezolid mitigates tissue injury in experimental model of pediatric testicular torsion: TLR-4/MAPK/NF-κB involvement
Background: Testicular torsion is a urological emergency that requires prompt surgery to prevent orchiectomy. Pharmacological interventions may slow the progression of damage and reduce reperfusion injury after surgical correction.Purpose: This study evaluated the protective effects of linezolid against testicular torsion-detorsion (T/D) injury in rats by focusing on the mechanisms involving the Toll-like receptor 4 (TLR-4) pathway.Methods: Eighty-four male Wistar rats were allocated into 8 groups; of them, one was subjected to a sham operation and another was subjected to 4-hour ischemia via 720° of torsion followed by 24-hour reperfusion. Linezolid (3–100 mg/kg) was assessed for its effects on T/D injury using histopathological evaluation, oxidative stress markers (malondialdehyde [MDA], superoxide dismutase [SOD]), and inflammatory biomarker tumor necrosis factor-alpha (TNF-α). Mechanistic investigations have focused on TLR-4 the mitogen-activated protein kinase (MAPK)/nuclear factor kappa B (NF-κB) pathway. Molecular docking and in silico analyses were conducted to predict interactions with key inflammatory proteins.Results: Linezolid 25, 50, and 100 mg/kg significantly reduced the histopathological damage, with 50 mg/kg being the most effective dosage. Within the 6–50 mg/kg range, linezolid reduced MDA, increased SOD, decreased TNF-α, and suppressed TLR-4/NF-κB pathway activity, with maximal reductions in MDA, TNF-α, NF-κB, and TLR-4 of 64%, 77%, 56%, and 53%, respectively, and an enhancement in SOD of 47%. In silico docking predicted strong binding interactions with TLR-4 pathway proteins, including p38 MAPK and JNK, with affinities of -7.4 to -8.3 kcal/mol.Conclusion: Linezolid protects against testicular torsion by reducing oxidative stress and inflammation via modulating the TLR-4/NF-κB pathway, suggesting its therapeutic potential and need for further study.
- 15 September 2025
- Published by Korean Pediatric Society in Clinical and Experimental Pediatrics
- Vol. 68 (9) , 700-711
Open AccessLinezolid mitigates tissue injury in experimental model of pediatric testicular torsion: TLR-4/MAPK/NF-κB involvement
Background: Testicular torsion is a urological emergency that requires prompt surgery to prevent orchiectomy. Pharmacological interventions may slow the progression of damage and reduce reperfusion injury after surgical correction.Purpose: This study evaluated the protective effects of linezolid against testicular torsion-detorsion (T/D) injury in rats by focusing on the mechanisms involving the Toll-like receptor 4 (TLR-4) pathway.Methods: Eighty-four male Wistar rats were allocated into 8 groups; of them, one was subjected to a sham operation and another was subjected to 4-hour ischemia via 720° of torsion followed by 24-hour reperfusion. Linezolid (3–100 mg/kg) was assessed for its effects on T/D injury using histopathological evaluation, oxidative stress markers (malondialdehyde [MDA], superoxide dismutase [SOD]), and inflammatory biomarker tumor necrosis factor-alpha (TNF-α). Mechanistic investigations have focused on TLR-4 the mitogen-activated protein kinase (MAPK)/nuclear factor kappa B (NF-κB) pathway. Molecular docking and in silico analyses were conducted to predict interactions with key inflammatory proteins.Results: Linezolid 25, 50, and 100 mg/kg significantly reduced the histopathological damage, with 50 mg/kg being the most effective dosage. Within the 6–50 mg/kg range, linezolid reduced MDA, increased SOD, decreased TNF-α, and suppressed TLR-4/NF-κB pathway activity, with maximal reductions in MDA, TNF-α, NF-κB, and TLR-4 of 64%, 77%, 56%, and 53%, respectively, and an enhancement in SOD of 47%. In silico docking predicted strong binding interactions with TLR-4 pathway proteins, including p38 MAPK and JNK, with affinities of -7.4 to -8.3 kcal/mol.Conclusion: Linezolid protects against testicular torsion by reducing oxidative stress and inflammation via modulating the TLR-4/NF-κB pathway, suggesting its therapeutic potential and need for further study.
- Open Access
- 7 January 2025
- Published by JMIR Publications Inc. in JMIR Preprints
Academic Influencers and X Social Network: A Mixed-Methods Analysis of The Impact on Academic Metrics, Communication, and Influence (Preprint)
BACKGROUND Background: Social media has redefined academic communication, with X (formerly Twitter) emerging as a central platform for sharing research, fostering collaboration, and expanding the reach of science. However, the connection between social media engagement and traditional academic metrics remains underexplored. OBJECTIVE Objective: This study evaluates the role of X in influencing academic impact metrics, focusing on the correlation between X engagement, Altmetric Attention Scores (AAS), citation counts, and the insights derived from the Narrative Review of Studies. METHODS Methods: Data from the top 100 Altmetric articles between 2017 and 2021 were analyzed. Altmetric scores, X mentions, and citation counts were assessed, and Pearson correlation analysis was conducted to explore year-by-year trends. The Narrative Review of Studies provides detailed contextual information, including the key characteristics of top influencers and engagement patterns across X. RESULTS Results: A strong positive correlation was observed between X mentions and AAS across all years, with coefficients ranging from 0.673 to 0.944 (p < 0.05). The correlation between AAS and citation counts was moderate and variable, peaking at 0.504 (p < 0.05) in 2020. The Narrative Review of Studies highlights the key factors influencing these metrics, such as the expertise and geographical concentration of influencers, and the diverse interaction patterns on X. These findings underscore the complementary role of X in boosting visibility and short-term engagement while showing a weaker link to long-term academic impact. CONCLUSIONS Conclusions: Social media, particularly X, enhances the visibility of academic work and complements traditional metrics by providing immediate insights into online engagement. While Altmetric scores and X activity increased visibility, they work best alongside established measures like citation counts to provide a comprehensive understanding of research impact. Further research on multi-platform engagement is needed to clarify the broader role of social media in academic dissemination.
- 7 January 2025
- Published by JMIR Publications Inc. in JMIR Preprints
Open AccessAcademic Influencers and X Social Network: A Mixed-Methods Analysis of The Impact on Academic Metrics, Communication, and Influence (Preprint)
BACKGROUND Background: Social media has redefined academic communication, with X (formerly Twitter) emerging as a central platform for sharing research, fostering collaboration, and expanding the reach of science. However, the connection between social media engagement and traditional academic metrics remains underexplored. OBJECTIVE Objective: This study evaluates the role of X in influencing academic impact metrics, focusing on the correlation between X engagement, Altmetric Attention Scores (AAS), citation counts, and the insights derived from the Narrative Review of Studies. METHODS Methods: Data from the top 100 Altmetric articles between 2017 and 2021 were analyzed. Altmetric scores, X mentions, and citation counts were assessed, and Pearson correlation analysis was conducted to explore year-by-year trends. The Narrative Review of Studies provides detailed contextual information, including the key characteristics of top influencers and engagement patterns across X. RESULTS Results: A strong positive correlation was observed between X mentions and AAS across all years, with coefficients ranging from 0.673 to 0.944 (p < 0.05). The correlation between AAS and citation counts was moderate and variable, peaking at 0.504 (p < 0.05) in 2020. The Narrative Review of Studies highlights the key factors influencing these metrics, such as the expertise and geographical concentration of influencers, and the diverse interaction patterns on X. These findings underscore the complementary role of X in boosting visibility and short-term engagement while showing a weaker link to long-term academic impact. CONCLUSIONS Conclusions: Social media, particularly X, enhances the visibility of academic work and complements traditional metrics by providing immediate insights into online engagement. While Altmetric scores and X activity increased visibility, they work best alongside established measures like citation counts to provide a comprehensive understanding of research impact. Further research on multi-platform engagement is needed to clarify the broader role of social media in academic dissemination.
- Open Access
- 2 June 2022
- Published by American Chemical Society (ACS) in ACS Omega
- Vol. 7 (23) , 19401-19411
Isoindolin-1-ones Fused to Barbiturates: From Design and Molecular Docking to Synthesis and Urease Inhibitory Evaluation
Helicobacter pylori-induced ulcers and gastric cancer have been one of the main obstacles that the human community has ever struggled with, especially in recent decades. Several different attempts have been made to eradicate this group. One of the most widely used attempts is to inhibit the critical enzyme that facilitates its survival, the urease enzyme. Therefore, in this study, isoindolin-1-ones fused to barbiturates were designed, synthesized, and evaluated for their in vitro urease inhibitory activity as novel inhibitors for the urease enzyme. The synthesis route consisted of two steps. These steps increased the yield rate and decreased the percentage of byproducts while approaching green chemistry using ethanol and water as green solvents and microwave irradiation instead of conventional methods. In vitro urease inhibitory results indicated that all the compounds had higher inhibitory activity than the standard inhibitor, thiourea, and compound 5b proved to be the most potent inhibitor (IC50 = 0.82 ± 0.03 μM). A molecular docking study was performed to understand the interaction between compounds 5a-n and Jack bean urease enzyme. The results of the molecular docking study were also in harmony with the in vitro results, which are discussed in detail later in this study.
- 2 June 2022
- Published by American Chemical Society (ACS) in ACS Omega
- Vol. 7 (23) , 19401-19411
Open AccessIsoindolin-1-ones Fused to Barbiturates: From Design and Molecular Docking to Synthesis and Urease Inhibitory Evaluation
Helicobacter pylori-induced ulcers and gastric cancer have been one of the main obstacles that the human community has ever struggled with, especially in recent decades. Several different attempts have been made to eradicate this group. One of the most widely used attempts is to inhibit the critical enzyme that facilitates its survival, the urease enzyme. Therefore, in this study, isoindolin-1-ones fused to barbiturates were designed, synthesized, and evaluated for their in vitro urease inhibitory activity as novel inhibitors for the urease enzyme. The synthesis route consisted of two steps. These steps increased the yield rate and decreased the percentage of byproducts while approaching green chemistry using ethanol and water as green solvents and microwave irradiation instead of conventional methods. In vitro urease inhibitory results indicated that all the compounds had higher inhibitory activity than the standard inhibitor, thiourea, and compound 5b proved to be the most potent inhibitor (IC50 = 0.82 ± 0.03 μM). A molecular docking study was performed to understand the interaction between compounds 5a-n and Jack bean urease enzyme. The results of the molecular docking study were also in harmony with the in vitro results, which are discussed in detail later in this study.
- 1 March 2019
- Published by Elsevier in Drug Discovery Today
- Vol. 24 (3) , 898-905
Fullerene-based delivery systems
- 1 March 2019
- Published by Elsevier in Drug Discovery Today
- Vol. 24 (3) , 898-905
Fullerene-based delivery systems
- 1 March 2019
- Published by Elsevier in Materials Chemistry and Physics
- Vol. 225, 331-339
Corrosion behavior and in-vitro bioactivity of porous Mg/Al2O3 and Mg/Si3N4 metal matrix composites fabricated using microwave sintering process
- 1 March 2019
- Published by Elsevier in Materials Chemistry and Physics
- Vol. 225, 331-339
Corrosion behavior and in-vitro bioactivity of porous Mg/Al2O3 and Mg/Si3N4 metal matrix composites fabricated using microwave sintering process
- 1 October 2018
- Published by Elsevier in Ceramics International
- Vol. 44 (15) , 18743-18751
Preparation of mullite/B4C composites: A comparative study on the effect of heating methods
- 1 October 2018
- Published by Elsevier in Ceramics International
- Vol. 44 (15) , 18743-18751
Preparation of mullite/B4C composites: A comparative study on the effect of heating methods