Quantum random number generation based on phase reconstruction.

Quantum random number generator (QRNG) utilizes the intrinsic randomness of quantum systems to generate completely unpredictable and genuine random numbers, finding wide applications across many fields. QRNGs relying on the phase noise of a laser have attracted considerable attention due to their straightforward system architecture and high random number generation rates. However, traditional phase noise QRNGs suffer from a 50% loss of quantum entropy during the randomness extraction process. In this paper, we propose a phase-reconstruction quantum random number generation scheme, in which the phase noise of a laser is reconstructed by simultaneously measuring the orthogonal quadratures of the light field using balanced detectors. This enables direct discretization of uniform phase noise, and the min-entropy can achieve a value of 1. Furthermore, our approach exhibits inherent robustness against the classical phase fluctuations of the unbalanced interferometer, eliminating the need for active compensation. Finally, we conducted experimental validation using commercial optical hybrid and balanced detectors, achieving a random number generation rate of 1.96 Gbps at a sampling rate of 200 MSa/s.

Effect of tannic acid modification on the interface and emulsification properties of zein colloidal particles.

BACKGROUND: Interface modification driven by supramolecular self-assembly has been accepted as a valuable strategy for emulsion stabilization enhancement. However, there has been a dearth of comparative research on the effect of simple complexation and assembly from the perspective of the responsible mechanism. RESULTS: The present study selected zein and tannic acid (TA) as representative protein and polyphenol modules for self-assembly (coined as TA-modified zein particle and TA-zein complex particle) to explore the surface properties and interfacial behavior, as well as the stability of constructed Pickering emulsions to obtain the regulation law of different modification methods on the interfacial behavior of colloidal particles. The results demonstrated that TA-modified zein colloidal particles potentially improved the emulsifying properties. When the TA concentration was 3 mmol L-1 , the optimized TA-modified zein particle was nano-sized (109.83 nm) and had advantageous interfacial properties, including sharply reduced surface hydrophobicity, as well as a low diffusion rate at the oil/water interface. As a result, the shelf life of Pickering emulsion containing 50% oil phase was extended to 90 days. CONCLUSION: Through multi-angled research on the properties of the interfacial membrane, improvement of emulsion stability was a result of the formation of viscoelastic interfacial film that resulted from the decrease of absorption rate between particles and interface. Using refined regulation to investigate the role of different sample preparation methods from a mechanistic perspective. Overall, the present study has provided a reference for TA to regulate the surface properties and interface behavior of zein colloidal particles, enriched the understanding of colloidal interface assembly, and provided a theoretical basis for the quality control of interface-oriented food systems. © 2023 Society of Chemical Industry.

Building Machine Learning Small Molecule Melting Points and Solubility Models Using CCDC Melting Points Dataset.

Predicting solubility of small molecules is a very difficult undertaking due to the lack of reliable and consistent experimental solubility data. It is well known that for a molecule in a crystal lattice to be dissolved, it must, first, dissociate from the lattice and then, second, be solvated. The melting point of a compound is proportional to the lattice energy, and the octanol-water partition coefficient (log P) is a measure of the compound's solvation efficiency. The CCDC's melting point dataset of almost one hundred thousand compounds was utilized to create widely applicable machine learning models of small molecule melting points. Using the general solubility equation, the aqueous thermodynamic solubilities of the same compounds can be predicted. The global model could be easily localized by adding additional melting point measurements for a chemical series of interest.

A Modified Single-Frequency PPP Method for the Positioning and Time Transfer with BDS-3.

In this paper, time-frequency transfer and positioning experiments with signal coexistence in the BDS system were conducted using the four types of open service signals of the BDS-3 satellite (B1I, B1C, B2a, and B3I), as well as the B2I signals broadcast by the BDS-2 satellites. The experiments used the single-frequency PPP (precise point positioning) method. The experiment validated a modified version of the group and phase ionospheric correction (GRAPHIC) technique. The results demonstrate that, with a single frequency, 18 selected stations may provide positioning results accurate to within a few decimeters. The positioning accuracy of five frequencies signals is improved by 40.4%, 32.2%, 80.3%, 12.4%, and 10.3% when compared to the positioning accuracy of the same signals when using the general observation approach. Currently, the frequency stability may be as precise as dual frequencies with BDS.

Collaborative Profile-QSAR: A Natural Platform for Building Collaborative Models among Competing Companies.

Massively multitask bioactivity models that transfer learning between thousands of assays have been shown to work dramatically better than separate models trained on each individual assay. In particular, the applicability domain for a given model can expand from compounds similar to those tested in that specific assay to those tested across the full complement of contributing assays. If many large companies would share their assay data and train models on the superset, predictions should be better than what each company can do alone. However, a company's compounds, targets, and activities are among their most guarded trade secrets. Strategies have been proposed to share just the individual collaborators' models, without exposing any of the training data. Profile-QSAR (pQSAR) is a two-level, multitask, stacked model. It uses profiles of level-1 predictions from single-task models for thousands of assays as compound descriptors for level-2 models. This work describes its simple and natural adaptation to safe collaboration by model sharing. Broad model sharing has not yet been implemented across multiple large companies, so there are numerous unanswered questions. Novartis was formed from several mergers and acquisitions. In principle, this should allow an internal simulation of model sharing. In practice, the lack of metadata about the origins of compounds and assays made this difficult. Nevertheless, we have attempted to simulate this process and propose some findings: multitask pQSAR is always an improvement over single-task models; collaborative multitask modeling did not improve predictions on internal compounds; collaboration did improve predictions for external compounds but far less than the purely internal multitask modeling for internal compounds; collaborative models for external compounds increasingly improve as overlap between compound collections increases; combining profiles from inside and outside the company is not best, with internal predictions better using only the inside profile and external using only the outside profile, but a consensus of models using all three profiles is best on external compounds and a good compromise on internal compounds. We anticipate similar results from other model-sharing approaches. Indeed, since collaborative pQSAR through model sharing is mathematically identical to pQSAR using actual shared data, we believe our conclusions should apply to collaborative modeling by any current method even including the unlikely scenario of directly sharing all chemical structures and assay data.

Electron-deficient TVT unit-based D-A polymer donor for high-efficiency thick-film OSCs.

As the power conversion efficiencies of organic solar cells (OSCs) have been improved continuously in recent years, more attention will be paid to the industrial production and practical application of OSCs. However, there are still many problems to be solved in the process of large-scale production. Among them, reducing the costs of the materials and enhancing the film-thickness tolerance of the active layer are the two key points. Therefore, it is urgent to develop organic semiconductor materials which are easy to synthesize and suitable for the construction of high-efficiency, thick-film OSCs. In this work, we have focused on the (E)-2-[2-(thiophen-2-yl)vinyl]thiophene (TVT) unit because of its unique coplanar structure. And we noticed that TVT was mostly used as an electron-donating unit in the previous reports. However, we have modified TVT into electron-withdrawing unit by the introduction of fluorine atoms/ester groups. And two new donor-acceptor (D-A) copolymers have been obtained by combining the electron-withdrawing TVT unit with benzo[2,1-b:4,5-b']dithiophene (BDT) unit. Among them, the polymer based on the ester modified TVT unit presents excellent photovoltaic performance by virtue of its good solubility and preferable molecular stacking mode, and the corresponding devices also show extraordinarily high-thickness tolerance. The emergence of this new electron-withdrawing TVT unit will undoubtedly further promote the development of low-cost, high-efficiency, thick-film OSCs.

Catechol-functionalized hydrogels: biomimetic design, adhesion mechanism, and biomedical applications.

Hydrogels are a unique class of polymeric materials that possess an interconnected porous network across various length scales from nano- to macroscopic dimensions and exhibit remarkable structure-derived properties, including high surface area, an accommodating matrix, inherent flexibility, controllable mechanical strength, and excellent biocompatibility. Strong and robust adhesion between hydrogels and substrates is highly desirable for their integration into and subsequent performance in biomedical devices and systems. However, the adhesive behavior of hydrogels is severely weakened by the large amount of water that interacts with the adhesive groups reducing the interfacial interactions. The challenges of developing tough hydrogel-solid interfaces and robust bonding in wet conditions are analogous to the adhesion problems solved by marine organisms. Inspired by mussel adhesion, a variety of catechol-functionalized adhesive hydrogels have been developed, opening a door for the design of multi-functional platforms. This review is structured to give a comprehensive overview of adhesive hydrogels starting with the fundamental challenges of underwater adhesion, followed by synthetic approaches and fabrication techniques, as well as characterization methods, and finally their practical applications in tissue repair and regeneration, antifouling and antimicrobial applications, drug delivery, and cell encapsulation and delivery. Insights on these topics will provide rational guidelines for using nature's blueprints to develop hydrogel materials with advanced functionalities and uncompromised adhesive properties.

All-Assay-Max2 pQSAR: Activity Predictions as Accurate as Four-Concentration IC50s for 8558 Novartis Assays.

Profile-quantitative structure-activity relationship (pQSAR) is a massively multitask, two-step machine learning method with unprecedented scope, accuracy, and applicability domain. In step one, a "profile" of conventional single-assay random forest regression models are trained on a very large number of biochemical and cellular pIC50 assays using Morgan 2 substructural fingerprints as compound descriptors. In step two, a panel of partial least squares (PLS) models are built using the profile of pIC50 predictions from those random forest regression models as compound descriptors (hence the name). Previously described for a panel of 728 biochemical and cellular kinase assays, we have now built an enormous pQSAR from 11 805 diverse Novartis (NVS) IC50 and EC50 assays. This large number of assays, and hence of compound descriptors for PLS, dictated reducing the profile by only including random forest regression models whose predictions correlate with the assay being modeled. The random forest regression and pQSAR models were evaluated with our "realistically novel" held-out test set, whose median average similarity to the nearest training set member across the 11 805 assays was only 0.34, comparable to the novelty of compounds actually selected from virtual screens. For the 11 805 single-assay random forest regression models, the median correlation of prediction with the experiment was only rext2 = 0.05, virtually random, and only 8% of the models achieved our standard success threshold of rext2 = 0.30. For pQSAR, the median correlation was rext2 = 0.53, comparable to four-concentration experimental IC50s, and 72% of the models met our rext2 > 0.30 standard, totaling 8558 successful models. The successful models included assays from all of the 51 annotated target subclasses, as well as 4196 phenotypic assays, indicating that pQSAR can be applied to virtually any disease area. Every month, all models are updated to include new measurements, and predictions are made for 5.5 million NVS compounds, totaling 50 billion predictions. Common uses have included virtual screening, selectivity design, toxicity and promiscuity prediction, mechanism-of-action prediction, and others. Several such actual applications are described.

A Robust Salty Water Adhesive by Counterion Exchange Induced Coacervate.

Counterion exchange of charged macromolecules has comprehensive implications in biological and synthetic systems such as protein function, biosignaling, ion conducting, and separation, but the correlation between the dynamic ion exchange, polyelectrolyte phase separation, and functionality remains elusive. Here, counterion exchange is exploited as a means to facilitate liquid-liquid phase separation and coacervates featuring higher stability and versatility compared with conventional complex coacervate. Self-coacervation of a cationic polyelectrolyte (polyamidoamine-epichlorohydrin, PAE-Cl) occurs in broader conditions when its original counter anion (Cl- ) is exchanged by bis(trifluoromethane-sulphonyl)imide anion (TFSI- ), as a result of TFSI- counter anions association instead of polyelectrolyte complexation. This coacervate is catechol-free, easy to prepare, and highlights robust wet adhesion strength on diverse submerged surfaces in salty water (pH = 3-11), as demonstrated by its versatile capability of in situ underwater gluing and repairing without any pre-immersive drying.

From Alloy-Like to Cascade Blended Structure: Designing High-Performance All-Small-Molecule Ternary Solar Cells.

Ternary blending strategy has been used to design and fabricate efficient organic solar cells by enhancing the short-circuit current density and the fill factor. In this manuscript, we report all-small-molecule ternary solar cells consisting of two compatible small molecules DR3TBDTT (M1) and DR3TBDTT-E (M2) as donors and PC71BM as acceptor. A transformation from an alloy-like model to a cascade model are first realized by designing a novel molecule M2. It is observed that after thermal and solvent vapor annealing M2 shifts from the mixed region to donor-acceptor (D-A) interfaces which ameliorates the charge transfer and recombination processes. The optimal ternary solar cells with 10% M2 exhibited a power conversion efficiency of 8.48% in the alloy-like model and 10.26% in the cascade model. The proposed working mechanisms are fully characterized and further supported by the density functional theory and atomistic molecular dynamics simulations. This provides an important strategy to design high-performance ternary solar cells which contains one molecule not only is compatible with the main donor molecule but also performs a preference to appear at the D-A interfaces hence builds cascade energy levels.

Contamination, source identification, and risk assessment of polycyclic aromatic hydrocarbons in the soils of vegetable greenhouses in Shandong, China.

The concentrations of 16 polycyclic aromatic hydrocarbons (PAHs) were analyzed in soil (n=196) and vegetable (n=30) collected from greenhouses, and also in the soil (n=27) collected from agriculture fields close to the greenhouses in Shandong Province, China. The total PAH concentration (∑16PAH) ranged from 152.2µg/kg to 1317.7µg/kg, within the moderate range in agricultural soils of China. Three-ring PAHs were the dominant species, with Phe (16.3%), Ace (13.1%), and Fl (10.5%) as the major compounds. The concentrations of low molecular weight (LMW ≤3 rings) PAHs were high in the east and north of Shandong, while the concentrations of high molecular weight (HMW ≥4 rings) PAHs were high in the south and west of the study area. The PAH level in soils in industrial areas (IN) was higher than those in transport areas (TR) and rural areas (RR). No significant difference in concentration of ∑16PAH and composition was observed in soils of vegetable greenhouses and field soils. PAH concentration exhibited a weakly positive correlation with alkaline nitrogen, available phosphorus in soil, but a weakly negative correlation with soil pH. However, no obvious correlation was observed between PAH concentration and organic matter of soil, or ages of vegetable greenhouses. ∑16PAH in vegetables ranged from 89.9µg/kg to 489.4µg/kg, and LMW PAHs in vegetables positively correlated with those in soils. The sources of PAHs were identified and quantitatively assessed through positive matrix factorization. The main source of PAHs in RR was coal combustion, while the source was traffic in TR and IN. Moreover, petroleum source, coke source, biomass combustion, or mixed sources also contributed to PAH pollution. According to Canadian soil quality guidelines, exposure to greenhouse soils in Shandong posed no risk to human health.

Recursive Random Forests Enable Better Predictive Performance and Model Interpretation than Variable Selection by LASSO.

Variable selection is of crucial significance in QSAR modeling since it increases the model predictive ability and reduces noise. The selection of the right variables is far more complicated than the development of predictive models. In this study, eight continuous and categorical data sets were employed to explore the applicability of two distinct variable selection methods random forests (RF) and least absolute shrinkage and selection operator (LASSO). Variable selection was performed: (1) by using recursive random forests to rule out a quarter of the least important descriptors at each iteration and (2) by using LASSO modeling with 10-fold inner cross-validation to tune its penalty λ for each data set. Along with regular statistical parameters of model performance, we proposed the highest pairwise correlation rate, average pairwise Pearson's correlation coefficient, and Tanimoto coefficient to evaluate the optimal by RF and LASSO in an extensive way. Results showed that variable selection could allow a tremendous reduction of noisy descriptors (at most 96% with RF method in this study) and apparently enhance model's predictive performance as well. Furthermore, random forests showed property of gathering important predictors without restricting their pairwise correlation, which is contrary to LASSO. The mutual exclusion of highly correlated variables in LASSO modeling tends to skip important variables that are highly related to response endpoints and thus undermine the model's predictive performance. The optimal variables selected by RF share low similarity with those by LASSO (e.g., the Tanimoto coefficients were smaller than 0.20 in seven out of eight data sets). We found that the differences between RF and LASSO predictive performances mainly resulted from the variables selected by different strategies rather than the learning algorithms. Our study showed that the right selection of variables is more important than the learning algorithm for modeling. We hope that a standard procedure could be developed based on these proposed statistical metrics to select the truly important variables for model interpretation, as well as for further use to facilitate drug discovery and environmental toxicity assessment.

Hybrid in silico models for drug-induced liver injury using chemical descriptors and in vitro cell-imaging information.

Drug-induced liver injury (DILI) is a major adverse drug reaction that accounts for one-third of post-marketing drug withdrawals. Several classifiers for human hepatotoxicity using chemical descriptors with limited prediction accuracies have been published. In this study, we developed predictive in silico models based on a set of 156 DILI positive and 136 DILI negative compounds for DILI prediction. First, models based on a chemical descriptor (CDK, Dragon and MOE) and in vitro cell-imaging endpoints [human hepatocyte imaging assay technology (HIAT) descriptors] were built using random forest (RF) and five-fold cross-validation procedure. Then three hybrid models were built using HIAT and a single type of chemical descriptors. Generally, the models based only on chemical descriptors were poor, with a correct classification rate (CCR) around 0.60 when the default threshold value (i.e. threshold = 0.50) was used. The hybrid models afforded a CCR of 0.73 with a specificity of 0.74 and a better true positive rate (sensitivity of 0.71), which is crucial in drug toxicity screening for the purpose of patient safety. The benefit of hybrid models was even more drastic when stricter classification thresholds were employed (e.g. CCR would be 0.83 when double thresholds (non-toxic < 0.40 and toxic > 0.60) were used for the hybrid model). We have developed rigorously validated hybrid models which can be used in virtual screening of lead compounds with potential hepatotoxicity. Our study also showed a chemical structure and in vitro biological data can be complementary in enhancing the prediction accuracy of human hepatotoxicity and can afford rational mechanistic interpretation.

Two-stage prediction of the effects of imidazolium and pyridinium ionic liquid mixtures on luciferase.

The predicted toxicity of mixtures of imidazolium and pyridinium ionic liquids (ILs) in the ratios of their EC50, EC10, and NOEC (no observed effect concentration) were compared to the observed toxicity of these mixtures on luciferase. The toxicities of EC50 ratio mixture can be effectively predicted by two-stage prediction (TSP) method, but were overestimated by the concentration addition (CA) model and underestimated by the independent action (IA) model. The toxicities of EC10 ratio mixtures can be basically predicted by TSP and CA, but were underestimated by IA. The toxicities of NOEC ratio mixtures can be predicted by TSP and CA in a certain concentration range, but were underestimated by IA. Our results support the use of TSP as a default approach for predicting the combined effect of different types of ILs at the molecular level. In addition, mixtures of ILs mixed at NOEC and EC10 could cause significant effects of 64.1% and 97.7%, respectively. Therefore, we should pay high attention to the combined effects in mixture risk assessment.

Advances in Biosensors for the Rapid Detection of Marine Biotoxins: Current Status and Future Perspectives.

Marine biotoxins (MBs), harmful metabolites of marine organisms, pose a significant threat to marine ecosystems and human health due to their diverse composition and widespread occurrence. Consequently, rapid and efficient detection technology is crucial for maintaining marine ecosystem and human health. In recent years, rapid detection technology has garnered considerable attention for its pivotal role in identifying MBs, with advancements in sensitivity, specificity, and accuracy. These technologies offer attributes such as speed, high throughput, and automation, thereby meeting detection requirements across various scenarios. This review provides an overview of the classification and risks associated with MBs. It briefly outlines the current research status of marine biotoxin biosensors and introduces the fundamental principles, advantages, and limitations of optical, electrochemical, and piezoelectric biosensors. Additionally, the review explores the current applications in the detection of MBs and presents forward-looking perspectives on their development, which aims to be a comprehensive resource for the design and implementation of tailored biosensors for effective MB detection.

Effects of Muscle Type and Aging on Glycolysis and Physicochemical Quality Properties of Bactrian camel (Camelus bactrianus) Meat.

Poor tenderness of camel meat has seriously hampered the development of the camel meat industry. This study investigated the effects of muscle fiber composition and ageing time on meat quality, glycolytic potential, and glycolysis-related enzyme activities. Muscle samples of the longissimus thoracis (LT), psoas major (PM), and semitendinosus (ST) were collected from eight 8-10 year old Sonid Bactrian camels (females). Muscle fiber composition was examined by ATPase staining and immunohistochemistry. Meat quality indexes, glycolytic potential, and activities of major glycolytic enzymes were examined at 4 °C aging for 1, 6, 24, 72, and 120 h. The results showed that LT was mainly composed of type IIb muscle fibers, whereas PM and ST were mainly composed of type I muscle fibers. The PCR results of the myosin heavy chain (MyHC) were consistent with the ATPase staining results. During aging, the shear force of LT muscle was always greater than that of PM and ST, and its glycolysis was the strongest; type IIa, IIb, and IIx muscle fibers were positively correlated with muscle shear force and glycolysis rate, and type I muscle fibers were significantly and negatively correlated with the activities of the key enzymes of glycolysis within 6 h. The results showed that the muscle fibers of LT muscle had the greatest glycolysis capacity. These results suggest that an excessive type IIb muscle fiber number percentage and area in camel meat accelerated the glycolysis process, but seriously affected the sensory profile of the camel meat. The results of this study provide directions for the camel industry when addressing the poor tenderness of camel meat.

Adenosine monophosphate boosts the cryoprotection of ultrasound-assisted freezing to frozen surimi: Insights into protein structures and gelling behaviors.

Ultrasound-assisted freezing (UAF) is a clean technique for meat cryoprotections; however, its effectiveness is still limited compared to conventional cryoprotectants, e.g., sugars, polyols, especially at high dosages. To resolve this problem, a synergistic cryoprotection strategy was developed in this study. Adenosine monophosphate (AMP), an adenosine-type food additive, was introduced into frozen surimi at a considerably reduced content (0.08%), yet substantially enhanced the efficiency of UAF to comparable levels of commercial cryoprotectant (4% sucrose with 4% sorbitol). Specifically, UAF/AMP treatment retarded denaturation of surimi myofibrillar protein (MP) during 60-day frozen storage, as evidenced by its increased solubility, Ca2+-ATPase activity, sulfhydryl content, declined surface hydrophobicity, particle size, and stabilized protein conformation. Gels of UAF/AMP-treated surimi also demonstrated more stabilized microstructures, uniform water distributions, enhanced mechanical properties and water-holding capacities. This study provided a feasible approach to boost the cryoprotective performance of UAF, thus expanding its potential applications in frozen food industry.

Metabolomic insights into the effect of chickpea protein hydrolysate on the freeze-thaw tolerance of industrial yeasts.

The use of frozen dough is an intensive food-processing practice that contributes to the development of chain operations in the bakery industry. However, the fermentation activity of yeasts in frozen dough can be severely damaged by freeze-thaw stress, thereby degrading the final bread quality. In this study, chickpea protein hydrolysate significantly improved the quality of steamed bread made from frozen dough while enhancing the yeast survival rate and maintaining yeast cell structural integrity under freeze-thaw stress. The mechanism underlying this protective role of chickpea protein hydrolysate was further investigated by untargeted metabolomics analysis, which suggested that chickpea protein hydrolysate altered the intracellular metabolites associated with central carbon metabolism, amino acid synthesis, and lipid metabolism to improve yeast cell freeze-thaw tolerance. Therefore, chickpea protein hydrolysate is a promising natural antifreeze component for yeast cryopreservation in the frozen dough industry.

Dual-emissive Eu(III)-functionalized metal-organic frameworks for visual, rapid, and intelligent sensing of albendazole and albendazole sulfoxide in animal-origin food.

Albendazole (ABZ), a benzimidazole-based anthelmintic, is widely used to treat helminth infections. The extensive and improper use of ABZ may cause drug residues in animal-origin food and anthelmintics resistance, which potentially threaten human health. Meanwhile, albendazole sulfoxide (ABZSO), a metabolite of ABZ, also exhibits toxic effects. Therefore, the detection of ABZ and ABZSO in animal-derived food is significantly necessary. Herein, a dual-emission europium fluorescent sensor (EuUHC-30) was rationally designed and constructed. EuUHC-30 exhibits high selectivity and sensitivity towards ABZ and ABZSO with a detection limit of 0.10 and 0.13 µM, respectively. Furthermore, EuUHC-30 was successfully applied for quantification of ABZ and ABZSO in milk and pig kidney, which were verified by HPLC analysis. Moreover, a smartphone-assisted EuUHC-30 fluorescent paper sensor was fabricated for the practical determination of ABZ and ABZSO in real food. Overall, this work provides a visual, rapid, and intelligent method for the detection of ABZ and ABZSO in animal-origin food.

The use of pseudo-equilibrium constant affords improved QSAR models of human plasma protein binding.

PURPOSE: To develop accurate in silico predictors of Plasma Protein Binding (PPB). METHODS: Experimental PPB data were compiled for over 1,200 compounds. Two endpoints have been considered: (1) fraction bound (%PPB); and (2) the logarithm of a pseudo binding constant (lnKa) derived from %PPB. The latter metric was employed because it reflects the PPB thermodynamics and the distribution of the transformed data is closer to normal. Quantitative Structure-Activity Relationship (QSAR) models were built with Dragon descriptors and three statistical methods. RESULTS: Five-fold external validation procedure resulted in models with the prediction accuracy (R²) of 0.67 ± 0.04 and 0.66 ± 0.04, respectively, and the mean absolute error (MAE) of 15.3 ± 0.2% and 13.6 ± 0.2%, respectively. Models were validated with two external datasets: 173 compounds from DrugBank, and 236 chemicals from the US EPA ToxCast project. Models built with lnKa were significantly more accurate (MAE of 6.2-10.7 %) than those built with %PPB (MAE of 11.9-17.6 %) for highly bound compounds both for the training and the external sets. CONCLUSIONS: The pseudo binding constant (lnKa) is more appropriate for characterizing PPB binding than conventional %PPB. Validated QSAR models developed herein can be applied as reliable tools in early drug development and in chemical risk assessment.