Targeting Properties Research Articles

Inverse design of promising electrocatalysts for CO2 reduction via generative models and bird swarm algorithm

Directly generating material structures with optimal properties is a long-standing goal in material design. Traditional generative models often struggle to efficiently explore the global chemical space, limiting their utility to localized space. Here, we present a framework named Material Generation with Efficient Global Chemical Space Search (MAGECS) that addresses this challenge by integrating the bird swarm algorithm and supervised graph neural networks, enabling effective navigation of generative models in the immense chemical space towards materials with target properties. Applied to the design of alloy electrocatalysts for CO2 reduction (CO2RR), MAGECS generates over 250,000 structures, achieving a 2.5-fold increase in high-activity structures (35%) compared to random generation. Five predicted alloys— CuAl, AlPd, Sn2Pd5, Sn9Pd7, and CuAlSe2 are synthesized and characterized, with two showing around 90% Faraday efficiency for CO2RR. This work highlights the potential of MAGECS to revolutionize functional material development, paving the way for fully automated, artificial intelligence-driven material design.

Leveraging Multi-Task Learning Regressor Chains for Small and Sparse Tabular Data in Materials Design

Abstract Machine learning has become increasingly important in materials design, yet traditional single-task learning (STL) models fail to fully exploit the potential of available data in scenarios involving multiple targets and incomplete datasets. While STL models overlook the inherent correlations between target properties, this study showcases how multi-task learning (MTL) effectively leverages these correlations. Therefore, the performance of MTL methods compared to single-task learning (STL) is evaluated across five datasets, covering twelve prediction tasks and incorporating different types and levels of data sparsity. Our findings reveal that MTL significantly outperforms STL, particularly in sparse data scenarios, with up to 15% prediction improvements across all tasks. Moreover, MTL methods utilizing regressor chains with automated machine learning tools achieve superior performance compared to those based on neural networks, with minimal training effort required. This work advances data efficiency in data-driven materials design, establishing MTL as a potent tool for simultaneous learning and predicting multiple material properties.

Breaking barriers in targeted Therapy: Advancing exosome Isolation, Engineering, and imaging.

Exosomes have emerged as promising tools for targeted drug delivery in biomedical applications and medicine. This review delves into the scientific advancements, challenges, and future prospects specifically associated with these technologies. In this work, we trace the research milestones that led to the discovery and characterization of exosomes and extracellular vesicles, and discuss strategies for optimizing the synthetic yield and the loading of these particles with various therapeutics. In addition, we report the current major issues affecting the field and hampering the clinical translation of these technologies. Highlighting the pivotal role of imaging techniques, we explore how they drive exosome therapy and development by offering insights into biodistribution and cellular trafficking dynamics. Methodologies for vesicle isolation, characterization, loading, and delivery mechanisms are thoroughly examined, alongside strategies aimed at enhancing their therapeutic efficacy. Special emphasis was dedicated to their therapeutic properties, particularly to their ability to deliver biologics into the cytoplasm. Furthermore, we delve into the intricate balance between surface modifications and targeting properties including also transgenic methods aimed at their functionalization and visualization within biological systems. This review underscores the transformative potential of these carriers in targeted drug delivery and identifies crucial areas for further research and clinical translation.

In Situ Gels for Nasal Delivery: Formulation, Characterization and Applications

AbstractThe nasal route offers many advantages for drug delivery: quick onset of action, better patient compliance, avoidance of first‐pass metabolism and bypassing the blood‐brain barrier. Despite the potential of this route, several challenges exist, such as the short drug retention time caused by mucociliary clearance (MCC). In situ forming gels, which undergo a sol‐to‐gel transition with specific triggers at the site of action, present real opportunities in this field. They combine the intrinsic characteristics of hydrogels (elasticity and water‐holding capacity) with responsiveness, allowing easy application of drugs (spraying or extruding through a nozzle), as well as prolonged retention in the nasal cavity. The incorporation of mucoadhesive polymers, additives and nanocarriers can further tune the properties of in situ gels as nasal delivery platforms. This review summarizes advances in in situ gels for nasal drug delivery. We first describe challenges of the nasal route, target properties of in situ nasal gels, and then present both “classic” gelling polymers (poloxamers and polysaccharides) – which form the bulk of reported studies on nasal gels – as well as novel bespoke materials; we review characterization methods, highlighting the lack of standardization and accepted target values, then discuss applications by spraying, and conclude with future prospects.

<b>Multitarget-Directed Multiple Ligands in Anti-VEGF Resistant Glioblastoma Therapeutics: An in Silico Approach to Identify Potential Phytochemicals</b>

Angiogenesis is an important process in tumor progression. Vascular Endothelial Growth Factor (VEGF) is the key factor regulating angiogenesis and hence anti-VEGF therapy is considered a useful therapeutic approach in tumor conditions. However, the drug resistance and lack of efficacy of existing drugs limit the potential of such a therapeutic approach in certain cases and the tumor growth will continue through alternative mechanisms. Glioblastoma (GBM) is one such type of tumor that shows resistance to anti-VEGF therapy. Previously, we identified the hub genes differentially expressed in anti-VEGF resistance in GBM. Medhya Rasayana, an ayurvedic formulation, is used for the management of neurological disorders. In the present study, we used computational docking methods to identify the phytochemicals present in the medicinal plants of Medhya Rasayana, which can target the proteins expressed by the hub genes associated with anti-VEGF resistance. Network pharmacological analysis was also performed to identify the highly effective phytochemicals for a possible adjuvant therapy. Results showed that multiple phytochemicals of Glycirrhiza glabra Linn, Evolvulus alsinoides and Celastrus paniculatus target the anti-VGEF resistant proteins in GBM. This indicates the multi targeting property of Medhya Rasayana plant – phytocompounds and hence it may be considered for adjuvant therapy along with anti-VEGF therapy. Received: 10 July 2024 | Revised: 11 October 2024 | Accepted: 6 January 2025 Conflicts of Interest The authors declare that they have no conflicts of interest to this work. Data Availability Statement The data that support this work are available upon reasonable request to the corresponding author. Author Contribution Statement Kesavan R. Arya: Methodology, Formal analysis, Investigation, Writing - original draft, Writing - review & editing, Visualization, Funding acquisition. Sasikumar J. Soumya: Methodology, Formal analysis, Investigation, Writing - original draft, Writing - review & editing,Visualization, Funding acquisition. Anuroopa G. Nadh: Formal analysis, Investigation, Funding acquisition. Thankamani R. Aswathy: Formal analysis, Investigation. Vijayalakshmi B.: Investigation. Achuthsankar S Nair: Resources, Project administration. Oommen V. Oommen: Validation, Resources, Writing - review & editing, Project administration. Perumana R. Sudhakaran: Conceptualization, Methodology, Validation, Investigation, Resources, Writing - review & editing, Project administration, Funding acquisition

N-acetylcysteine and chitosan conjugate modified dexamethasone nanostructured lipid carriers: Enhanced permeability, precorneal retention and lower inflammation for the treatment of dry eye syndrome.

Dexamethasone (Dex) is a primary medication for treating dry eye syndrome, and tobramycin-dexamethasone eye drops are commercially available. However, the eye's complex physiological environment reduces its bioavailability, and repeated use can lead to significant systemic toxicity and side effects. This study introduces a novel conjugate of chitosan (CS) and N-acetylcysteine (NAC), a bioadhesive material, which was grafted onto the surface of a Dex-supported nanostructured lipid carrier (NLC) to develop an innovative nanoparticle lipid ocular drug delivery system (CS-NAC@Dex-NLC). The enhancements afforded by CS-NAC, such as adhesion, osmotic, and targeting properties, address the limitations of traditional eye drops. NMR characterization confirmed the successful synthesis of the copolymer (CS-NAC). Particle size (PS), zeta potential, and transmission electron microscopy (TEM) verified the proper formation of the CS-NAC@Dex-NLC system. Cytotoxicity tests confirmed its excellent biocompatibility and safety. Cellular uptake studies showed that CS-NAC@Dex-NLC achieved the highest efficiency. Pharmacokinetic assessments revealed a significant increase in Dex's bioavailability in tears and aqueous humor. In vitro corneal penetration and in vivo imaging experiments demonstrated effective corneal penetration and retention, enhancing the drug's duration on the ocular surface and overcoming the barrier effect. Pharmacodynamic studies in rabbits with dry eye syndrome indicated that CS-NAC@Dex-NLC effectively alleviates symptoms, repairs corneal damage, and stabilizes the tear film. ELISA results showed a reduction in inflammation caused by dry eye. These findings suggest that CS-NAC@Dex-NLC is a promising vector for dry eye treatment, offering significant clinical relevance.

Effect of probiotic extracellular vesicles and their applications on health and disease.

Probiotics have been established to exert a positive impact on the treatment of various diseases. Indeed, these active microorganisms have garnered significant attention in recent years for their potential to prevent and treat illnesses. Their beneficial effects have been hypothesized to be linked to their released extracellular vesicles. These nanoscale structures, secreted during the growth and metabolism of probiotics, possess favorable biocompatibility and targeting properties, thereby promoting intercellular material transport and signaling. This article aimed to review the bioactive components and functions of these probiotics vesicles, highlighting their role in the treatment of various diseases and discussing their potential future applications. By exploring the mechanisms of probiotic extracellular vesicles in disease development, this review aimed to provide a theoretical reference for further research on their therapeutic potential. © 2025 Society of Chemical Industry.

Monosaccharide coatings on nanoparticles affect protein corona formation but not the interaction with their binding receptor

Surface coatings with polyethylene glycol (PEG) polymers have often been employed to improve nanoparticles (NPs) biocompatibility and extend circulation time by reducing protein adsorption. PEGylated NPs benefit from steric hindrance and repulsion effects, which are influenced by PEG molecular weight, density, and chain conformation. However, repetitive exposure to PEG can trigger acute and chronic immunological responses as a result of the development of Immunoglobulin G anti-PEG antibodies. NPs functionalisation with glycans has become an emerging approach to increase their biocompatibility as these biomolecules are highly hydrophilic, biocompatible interact with biological receptors expressed in the body, and can be conjugated, controlling their orientation. In this study, we developed a series of gold NPs (AuNPs) coated with PEG linkers of different lengths and conjugated with mannose (Man) or sialic acid (Sia) glycans, and we carried out a detailed characterisation prior to and after exposure to biological fluids to study their behaviour and protein corona formation. Our findings show that the glycan-coated NPs exhibit stabilisation after protein interaction, with Man coatings showing the lowest protein affinity and that the glycans are biologically active and capable of binding to glycan receptors (such as Concanavalin A) despite the presence of a complex protein environment. Results indicate that glycan modification of PEGylated NPs reduces nonspecific interactions while preserving active targeting properties, underscoring their potential for therapeutic applications.

Modern state of the problem of liposomal complex laser phoresis in the treatment of patients with peri-implantitis

Peri-implantitis is a common problem, leading to tissue destruction and implant loss. The method of treatment for peri-implantitis should be selected depending on the extent of the disease. The diverse physicochemical and targeting properties of nanodelivery systems provide a favorable platform for drug delivery for the treatment of peri-implantitis. The review presented herein analyzes domestic and foreign literature devoted to the issues of possibility and problems of application of laser phoresis and liposomal complex in the treatment of peri-implantitis patients. The analysis and systematization of available modern data on the application of laser technologies and liposomal complexes in periodontal treatment is carried out. The absence of clinical researches on this subject with the minimal level of studying the use of this technique in the treatment of periodontal pathologies is established, which determines the relevance of further researches in this direction in order to provide more effective rendering of dental implantation services.

Development of polarizable and hydration-focused water models for the Martini 3 force field

Coarse-grained simulations have been increasingly employed in many fields of molecular chemistry. These models enable the evaluation of complex molecular structures by simplifying their description compared to all-atoms models. Specifically, the Martini 3 force field has been developed to describe biomolecular structures such as proteins, lipids, and carbohydrates. Their “building block” approach has been successfully used in modeling complex systems ranging from carbohydrate solutions, biological membranes, and even deep eutectic solvents. However, one main issue arises when considering interfacial systems: water is represented as a single particle with no effect of electrostatic interactions. In this work, we developed a polarizable water model for the Martini 3 force field, introducing charged particles to model the electrostatic interactions present in water systems accurately. Gaussian Processes were trained with simulation data for both bulk water and cross-interaction parameterization to consistently explore the parameter space and fine-tune our model to the target properties. The developed model showed good performance in reproducing the density and dielectric constant of bulk water. Regarding cross-interaction with other beads, two different sets of parameters were developed. The first, called the polarizable model, is parameterized to reproduce the hydration free energies of the original Martini 3 organic beads. This allows for a good description of free energies of transfer between water and organic phases, the main target property for the Martini 3 force field. The second, called the polarizable-hydration model, is parameterized to match the experimental hydration free energy of selected molecules. This allows for more accurate evaluation of hydration properties, which can be useful for water solutions. In particular, we show that the partitioning and interfacial behavior of nonionic surfactants is improved by the polarizable-hydration model when compared to the standard Martini 3 water model.

Handling non-linearities and pre-processing in multivariate calibration of vibrational spectra

Backgroundseveral automatic pre-processing methods are available for coping with scattering effects in vibrational spectra (near and mid infrared, Raman) when linear multivariate models are applied for calibration. In contrast, almost no methods have been developed for data sets showing a non-linear relationship between instrumental signals and analyte concentrations or target properties. Resultsseveral data sets are studied, both simulated and experimental, showing different degrees of non-linearity: very low, intermediate, and high. For each of them, three multivariate models have been applied: classical partial least-squares (PLS), sequential pre-processing through orthogonalization (SO-PLS) and Kernel partial least-squares (K-PLS). The success of these models depends on the degree of non-linearity presented by each data set. The results are discussed in terms of the model structures. Significancethe need of developing new automatic pre-processing techniques for non-linear vibrational data sets is highlighted, in order to reach a similar status as the linear counterparts.

Improving Molecular Design with Direct Inverse Analysis of QSAR/QSPR Model.

Recent advances in machine learning have significantly impacted molecular design, notably the molecular generation method combining the chemical variational autoencoder (VAE) with Gaussian mixture regression (GMR). In this method, a mathematical model is constructed with X as the latent variable of the molecule and Y as the target properties and activities. Through direct inverse analysis of this model, it is possible to generate molecules with the desired target properties. However, this approach outputs many strings that do not follow the simplified molecular input line entry system grammar and generates unrealistic chemical structures in which the properties and activity do not satisfy the target values. In this study, we focus on hierarchical VAE using molecular graphs to address these issues. We confirm that the combination of hierarchical VAE and GMR does not generate invalid outputs and returns molecules that simultaneously satisfy multiple target values. Moreover, we use this method to identify several molecules that are predicted to exhibit activity against drug targets.

PSP-GEN: Stochastic inversion of the Process-Structure-Property chain in materials design through deep, generative probabilistic modeling

Inverse material design is a cornerstone challenge in materials science, with significant applications across many industries. Traditional approaches that invert the structure–property (SP) linkage to identify microstructures with targeted properties often overlook the feasibility of production processes, leading to microstructures that may not be manufacturable. Achieving both desired properties and a realizable manufacturing procedure necessitates inverting the entire Process-Structure-Property (PSP) chain. However, this task is fraught with challenges, including stochasticity along the whole modeling chain, the high dimensionality of microstructures and process parameters, and the inherent ill-posedness of the inverse problem. This paper proposes a novel framework, named PSP-GEN, for the goal-oriented material design that effectively addresses these challenges by modeling the entire PSP chain with a deep generative model. It employs two sets of continuous, microstructure- and property-aware, latent variables, the first of which provides a lower-dimensional representation that captures the stochastic aspects of microstructure generation, while the second is a direct link to processing parameters. This structured, low-dimensional embedding not only simplifies the handling of high-dimensional microstructure data but also facilitates the application of gradient-based optimization techniques. The effectiveness and efficiency of this method are demonstrated in the inverse design of two-phase materials, where the objective is to design microstructures with target effective permeability. We compare state-of-the-art alternatives in challenging settings involving limited training data, target property regions for which no training data is available, and design tasks where the process parameters and microstructures have high-dimensional representations.

Web-BO: towards increased accessibility of Bayesian optimisation (BO) for chemistry.

Historically, the chemical discovery process has predominantly been a matter of trial-and-improvement, where small modifications are made to a chemical system, guided by chemical knowledge, with the aim of optimising towards a target property or combination of properties. While a trial-and-improvement approach is frequently successful, especially when assisted by the help of serendipity, the approach is incredibly time- and resource-intensive. Complicating this further, the available chemical space that could, in theory, be explored is remarkably vast. As we are faced with near infinite possibilities and limited resources, we require improved search methods to effectively move towards desired optima, e.g. chemical systems exhibiting a target property, or several desired properties. Bayesian optimisation (BO) has recently gained significant traction in chemistry, where within the BO framework, prior knowledge is used to inform and guide the search process to optimise towards desired chemical targets, e.g. optimal reaction conditions to maximise yield, or optimal catalyst exhibiting improved catalytic activity. While powerful, implementing BO algorithms in practice is largely limited to interfacing via various APIs - requiring advanced coding experience and bespoke scripts for each optimisation task. Further, it is challenging to seamlessly link these with electronic lab notebooks via a graphical user interface (GUI). Ultimately, this limits the accessibility of BO algorithms. Here, we present Web-BO, a GUI to support BO for chemical optimisation tasks. We demonstrate its performance using an open source dataset and associated emulator, and link the platform with an existing electronic lab notebook, datalab. By providing a GUI-based BO service, we hope to improve the accessibility of data-driven optimisation tools in chemistry; https://suprashare.rcs.ic.ac.uk/web-bo/.

Homotypic cell membrane-camouflaged biomimetic PLGA nanoparticle loading triptolide for the treatment of hepatocellular carcinoma

Hepatocellular carcinoma (HCC) is the fourth leading cause of cancer-associated death worldwide. Beside early detection, early diagnosis, and early surgery, it is urgent to try new strategies for the treatment of HCC. Triptolide (TPL) has been employed to treat HCC. However, its clinical applications were restricted by the narrow therapeutic window, severe toxicity, and poor water-solubility. In this study, we developed cancer cell membrane-camouflaged biomimetic PLGA nanoparticles loading TPL (TPL@mPLGA) with the homologous targeting property for the treatment of HCC. The TPL@mPLGA was successfully prepared with particle size of 195.5 ± 7.5 nm and zeta potential at −21.5 ± 0.2 mV with good stability. The drug loading (DL) of TPL@mPLGA was 2.94%. After Huh-7 cell membrane coating, the natural Huh-7 cell membrane proteins were found to be retained on TPL@mPLGA, thus endowing the TPL@mPLGA with enhanced accumulation at tumor site, and better anti-tumor activity in vitro and in vivo when compared with TPL or TPL@PLGA. The TPL@mPLGA showed enhanced anti-tumor effects and reduced toxicity of TPL, which could be adopted for the treatment of HCC.

ReLMM: Reinforcement Learning Optimizes Feature Selection in Modeling Materials.

A challenge to materials discovery is the identification of the physical features that are most correlated to a given target material property without redundancy. Such variables necessarily comprise the optimal search domain in subsequent material design. Here, we introduce a reinforcement learning-based material model (ReLMM) as a tool for analyzing a given database in identifying a minimal or near minimal subset of physical features for the design of a material with a given target property. We aim for minimality in the selected subset with respect to its size─smaller being better─ while maintaining the desired accuracy of the prediction. We have shown, using synthetic multiscale data sets, that ReLMM can identify the relative importance of features, and thus help identify which should be selected across scales. In the context of semiconducting materials, ReLMM can be used to improve the prediction of the band gap by identifying which features should be selected in model building. For metal halide perovskites, ReLMM was seen to find a near minimal data set at least as well as, if not better than, state-of-the-art feature selection tools such as LASSO and XGBoost. We also found that our domain-science oriented approach can be used to uncover the hierarchical structure of a material from a database consisting of molecular-scale, mesoscale and device-scale features and labels in complementarity with an earlier hierarchical model called NestedAE.

Thermoelectric Material Performance (zT) Predictions with Machine Learning.

Research efforts using the tools in machine- and deep learning models have begun to show success in predicting target properties such as thermoelectric (TE) properties, including the figure of merit (zT). These models were trained on various data sources that used experimental, crystallographic, and density functional theory (DFT) data. We developed an interpretable model on a huge experimental data set of ∼160,000 data points to predict the performance of thermoelectric materials. The model predicts the results of three different test sets with high accuracy, such as the root-mean-square error (RMSE) ranging from 0.15 to 0.20 and the evaluation coefficients (R2) ranging from 0.80 to 0.67. Furthermore, we highlight probable reasons such as literature error, varied synthesis routes for the same material, different forms of crystallinity and morphology, and different particle sizes and densities for the deviation of predicted zT from the experimental zT results of the test sets. Lastly, using an experimental data set, our study is one of the few examples that predict a complex zT property directly across the entire gamut of TE materials.

Elastic moduli and thermal conductivity of quantum materials at finite temperature

We describe a theoretical and computational approach to calculate the vibrational, elastic, and thermal properties of materials from the low-temperature quantum regime to the high-temperature anharmonic regime. This approach is based on anharmonic lattice dynamics and the Boltzmann transport equation. It relies on second and third-order force constant tensors estimated by fitting temperature-dependent empirical potentials from path-integral quantum simulations with a first-principles machine learning Hamiltonian. The temperature-renormalized harmonic force constants are used to calculate the elastic moduli and the phonon modes of materials. Harmonic and anharmonic force constants are combined to solve the phonon Boltzmann transport equation to compute the lattice thermal conductivity. We demonstrate the effectiveness of this approach on bulk crystalline silicon in the temperature range from 50 to 1200 K, showing substantial improvement in the prediction of the temperature dependence of the target properties compared to experiments.

Unlocking Nature's Potential: Ferritin as a Universal Nanocarrier for Amplified Cancer Therapy Testing via 3D Microtissues.

In the existing development of extensive drug screening models, 3D cell cultures outshine conventional 2D monolayer cells by closely imitating the in vivo tumor microenvironment. This makes 3D culture a more physiologically relevant and convenient system in the regime of preclinical drug testing. In the nanomedicinal world, nanoconjugates as nanocarriers are largely hunted due to their capability of precisely binding to target cells and distributing essential dosages of therapeutic drugs with enhanced safety profiles. Thus, for boosted drug availability, the evolution from conventional drug treatment to combination therapies and last switching to drug carriers has gained significant progression in cancer cure. In contrast to conventional engineered nanoparticles, herein, we successfully designed biomolecule (ferritin)-based drug nanoconjugates effective both as a single drug (valproic acid-VPA) and twin-drug (valproic acid/doxorubicin-Dox) carriers, which dramatically enhance the proficiency of the tumor therapeutic modality. To question the reported adjuvant drug property of VPA, we progressed utilizing at first VPA alone as an effective yet exclusive tumor therapy when delivered via some carrier molecule, in particular protein. Subsequently, we paralleled this comprehensive investigation output to compare and test the coloading strategy of drugs and observe the synergistic and/or additive behavior of VPA in conjugation with other anticancer agents (Dox) while given via a carrier molecule. To approach this, VPA and/or Dox molecules were encapsulated into the ferritin (F) cavity using a thermosensitive synthesis method by maintaining the temperature at 60 °C. The successful encapsulation of drugs in the protein nanocage was confirmed through various characterization techniques. The F-VPA/F-VPA-Dox nanoconjugates exhibited similar morphology and structural characteristics to the hollow ferritin cage and showed significant cytotoxicity than the naked drugs when tested on physiologically relevant 3D spheroid models. Precisely, our first designed carrier nanoconjugate, i.e., F-VPA, offered more than a 3-fold increased intratumoral drug concentration than free VPA and significantly suppressed tumor growth after a single-dose treatment. However, our second modeled carrier nanoconjugate, viz. F-VPA-Dox, revealed an extended median survival period and lesser toxicity when administered at a much more effective dose (∼3-5 μM), in 3D tumor spheroid models of various cancer cell lines. All in all, importantly, ferritin nanoconjugates exhibited an enhanced tumor inhibition rate with a single-dose treatment, which further confirms the benefits of the active targeting property of these nanocarriers. Moreover, these nanocarriers also offer to deliver a significant dose of the therapeutic drug into tumor cells, alongside tremendous biocompatibility and safety profiles in numerous tumor 3D spheroid models.

Human-in-the-loop active learning for goal-oriented molecule generation

Machine learning (ML) systems have enabled the modelling of quantitative structure–property relationships (QSPR) and structure-activity relationships (QSAR) using existing experimental data to predict target properties for new molecules. These property predictors hold significant potential in accelerating drug discovery by guiding generative artificial intelligence (AI) agents to explore desired chemical spaces. However, they often struggle to generalize due to the limited scope of the training data. When optimized by generative agents, this limitation can result in the generation of molecules with artificially high predicted probabilities of satisfying target properties, which subsequently fail experimental validation. To address this challenge, we propose an adaptive approach that integrates active learning (AL) and iterative feedback to refine property predictors, thereby improving the outcomes of their optimization by generative AI agents. Our method leverages the Expected Predictive Information Gain (EPIG) criterion to select additional molecules for evaluation by an oracle. This process aims to provide the greatest reduction in predictive uncertainty, enabling more accurate model evaluations of subsequently generated molecules. Recognizing the impracticality of immediate wet-lab or physics-based experiments due to time and logistical constraints, we propose leveraging human experts for their cost-effectiveness and domain knowledge to effectively augment property predictors, bridging gaps in the limited training data. Empirical evaluations through both simulated and real human-in-the-loop experiments demonstrate that our approach refines property predictors to better align with oracle assessments. Additionally, we observe improved accuracy of predicted properties as well as improved drug-likeness among the top-ranking generated molecules.Scientific contributionWe present an adaptable framework that integrates AL and human expertise to refine property predictors for goal-oriented molecule generation. This approach is robust to noise in human feedback and ensures that navigating chemical space with human-refined predictors leverages human insights to identify molecules that not only satisfy predicted property profiles but also score highly on oracle models. Additionally, it prioritizes practical characteristics such as drug-likeness, synthetic accessibility, and a favorable balance between exploring diverse chemical space and exploiting similarity to existing training data.