SEARCH · Search NASA

Results for “substructuring”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,099 records · Page 61

All sources NASA NTRS DOE OSTI

DOE OSTI2023

Taming the Pseudoelastic Response of Nitinol Using Ion Implantation

Implantation of Ni 50.5 Ti 49.5 wire with 30 MeV Ni 6+ ions at doses (< 0.1 DPA) typically smaller than employed in the literature is shown to systematically alter the pseudoelastic response, with extrema in Berkovich nanoindentation load (+50%), hysteresis (–60%), and recoverable displacement (–19%) occurring at ~ 3.6 μm below the implantation surface. These extraordinary values are attributed to ~ 10 to 20 nm amorphous clusters that constrain the stress-induced B2-B19' phase transformation. This is substantiated by phase field simulations of crystalline-amorphous composites and molecular dynamics simulations of crystalline-vacancy cluster composites showing the spatial refinement of martensite caused by nm-scale defects. Finally, the results suggest that ion implantation may potentially expand the processing and performance space for NiTi, by creating amorphous defects at smaller length scales than dislocation substructures produced by conventional deformation processing.

36 MATERIALS SCIENCE↗

DOE OSTI2025

New layered quaternary Zintl pnictide oxides Ba 2 Zn 2 Pn 2 O ( Pn = Sb, Bi): Discovery, crystal structures, band engineering, and transport properties

Three new heteroanionic oxypnictides, Ba 2 Zn 2 Sb 2 O, Ba 2 Zn 2 Bi 2 O, and the solid solution Ba 2 Zn 2 Sb 2−x Bi x O (x ≈ 1.1–1.6), have been synthesized and structurally characterized. They are isostructural with their Mn-bearing analog, adopting the Ba 2 Mn 2 Sb 2 O-type structure (space group P6 3 /mmc, No. 194), and feature a double-layered 2D $^{2}_{∞}$ [Zn 2 Pn 2 O] 2- substructure (Pn = Sb, Bi, Sb/Bi) composed of corner-sharing, distorted tetrahedral ZnPn 3 O units. Electronic structure calculations reveal a systematic progression from semiconducting Ba 2 Zn 2 Sb 2 O to metallic Ba 2 Zn 2 Bi 2 O as Bi content increases. These trends are corroborated by transport property measurements, with Ba 2 Zn 2 Sb 0.9(1) Bi 1.1 O exhibiting relatively low electrical resistivity, high Hall mobilities of ∼160 cm 2 /V·s, and large Seebeck coefficients from 69 to 132 μV K −1 over the 300–600 K temperature range. Comparison with structurally related Zintl pnictides, such as SrIn 2 As 2 and PrZn 3 As 3 phases, situates Ba 2 Zn 2 Pn 2 O (Pn = Sb, Bi) within a broader family of heteroanionic oxypnictide Zintl compounds, highlighting their structural flexibility and amenability to band engineering. Finally, electronic structure and bonding considerations point to tunable semiconducting behavior and underscore the relevance of these materials for thermoelectric and topological applications.

Band engineering↗

DOE OSTI2015

The vaccinia virus E6 protein influences virion protein localization during virus assembly

Vaccinia virus mutants in which expression of the virion core protein gene E6R is repressed are defective in virion morphogenesis. E6 deficient infections fail to properly package viroplasm into viral membranes, resulting in an accumulation of empty immature virions and large aggregates of viroplasm. We have used immunogold electron microscopy and immunofluorescence confocal microscopy to assess the intracellular localization of several virion structural proteins and enzymes during E6R mutant infections. We find that during E6R mutant infections virion membrane proteins and virion transcription enzymes maintain a normal localization within viral factories while several major core and lateral body proteins accumulate in aggregated virosomes. The results support a model in which vaccinia virions are assembled from at least three substructures, the membrane, the viroplasm and a “pre-nucleocapsid”, and that the E6 protein is essential for maintaining proper localization of the seven-protein complex and the viroplasm during assembly. - Highlights: • Mutation of E6 disrupts association of viral membranes with viral core proteins • Mutation of E6 does not perturb viral membrane biosynthesis • Mutation of E6 does not perturb localization of viral transcription enzymes • Mutation of E6 causes mis-localization and aggregation of viral core proteins • Vaccinia assembly uses three subassemblies: membranes, viroplasm, prenucleocapsid.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2013

Identification of signals that facilitate isoform specific nucleolar localization of myosin IC

Myosin IC is a single headed member of the myosin superfamily that localizes to the cytoplasm and the nucleus, where it is involved in transcription by RNA polymerases I and II, intranuclear transport, and nuclear export. In mammalian cells, three isoforms of myosin IC are expressed that differ only in the addition of short isoform-specific N-terminal peptides. Despite the high sequence homology, the isoforms show differences in cellular distribution, in localization to nuclear substructures, and in their interaction with nuclear proteins through yet unknown mechanisms. In this study, we used EGFP-fusion constructs that express truncated or mutated versions of myosin IC isoforms to detect regions that are involved in isoform-specific localization. We identified two nucleolar localization signals (NoLS). One NoLS is located in the myosin IC isoform B specific N-terminal peptide, the second NoLS is located upstream of the neck region within the head domain. We demonstrate that both NoLS are functional and necessary for nucleolar localization of specifically myosin IC isoform B. Our data provide a first mechanistic explanation for the observed functional differences between the myosin IC isoforms and are an important step toward our understanding of the underlying mechanisms that regulate the various and distinct functions of myosin IC isoforms. - Highlights: ► Two NoLS have been identified in the myosin IC isoform B sequence. ► Both NoLS are necessary for myosin IC isoform B specific nucleolar localization. ► First mechanistic explanation of functional differences between the isoforms.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2013

Imaging Nicotine in Rat Brain Tissue by Use of Nanospray Desorption Electrospray Ionization Mass Spectrometry

Imaging mass spectrometry offers simultaneous detection of drugs, drug metabolites and endogenous substances in a single experiment. This is important when evaluating effects of a drug on a complex organ system such as the brain, where there is a need to understand how regional drug distribution impacts function. Nicotine is an addictive drug and its action in the brain is of high interest. Here we use nanospray desorption electrospray ionization, nano-DESI, imaging to discover the localization of nicotine in rat brain tissue after in vivo administration of nicotine. Nano-DESI is a new ambient technique that enables spatially-resolved analysis of tissue samples without special sample pretreatment. We demonstrate high sensitivity of nano-DESI imaging that enables detection of only 0.7 fmole nicotine per pixel in the complex brain matrix. Furthermore, by adding deuterated nicotine to the solvent, we examined how matrix effects, ion suppression, and normalization affect the observed nicotine distribution. Finally, we provide preliminary results suggesting that nicotine localizes to the hippocampal substructure called dentate gyrus.

dentate gyrus↗

DOE OSTI2020

Sensitive Top-Down Proteomics Analysis of a Low Number of Mammalian Cells Using a Nanodroplet Sample Processing Platform

Top-down proteomics is a powerful tool for characterizing genetic variations and post-translational modifications at intact protein level. That being said, one significant technical gap of top-down proteomics is the inability to analyze low amount of biological samples, which limits its access to isolated rare cells, fine-needle aspiration biopsies, and tissue substructures. Herein, we developed an ultrasensitive top-down platform by incorporating a microfluidic sample preparation system, termed nanoPOTS (Nanodroplet Processing in One pot for Trace Samples), into a top-down proteomics workflow. A unique combination of a nonionic detergent dodecyl-ß-D-maltopyranoside (DDM) with urea as protein extraction buffer significantly improved both protein extraction efficiency and sample recovery. We hypothesize that the DDM detergent improves protein recovery by efficiently reducing non-specific adsorption of intact proteins on container surfaces, while urea serves as a strong denaturant to disrupt noncovalent complexes and release intact proteins for downstream analysis. The nanoPOTS-based top-down platform reproducibly and quantitatively identified ~170 to ~620 proteoforms from ~70 to ~770 HeLa cells containing ~10 to ~115 ng of total protein. A variety of post-translational modifications including acetylation, myristoylation, and iron binding were identified using only less than 800 cells. We anticipate the nanoPOTS top-down proteomics platform will be broadly applicable in biomedical researches, particularly where clinical specimens are not available in amounts amenable to standard workflows.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2025

Increasing the Scale of the Mass Spectrometry Query Language Compendium with Explainable AI

A significant bottleneck in metabolomics data interpretation is the effective use of domain knowledge to assign structural information based on fragmentation patterns. The mass spectrometry query language (MassQL) aims to make this process accessible and applicable across multiple analysis platforms. While advanced computational methods are capable of predicting compound structures from fragmentation data, AI/ML approaches often rely on complex, opaque criteria that are difficult to interpret or modify. As a result, their predictive patterns cannot be readily translated into human-readable rules, such as those used in MassQL. Here, in this study, we introduce ChemEcho, a machine learning embedding method that converts tandem mass spectrometry data into sparse feature vectors containing peak and neutral mass subformulae to enhance explainable AI/ML-based methods. An advantage of this approach is that decision trees trained using these feature vectors can be directly translated to MassQL. Using a battery of decision trees trained using ChemEcho embeddings to predict molecular attributes, we generated over 1500 MassQL queries for 765 molecular features and evaluated their precision and recall. From these queries, the 50 highest-performing queries were integrated into the MassQL compendium. This set of generated MassQL queries included environmentally and biologically relevant classes such as PFAS and molecules containing phosphate or sulfate substructures. To illustrate the impact these queries would have on a typical metabolomics experiment, these MassQL queries were applied to a public metabolomics data set─resulting in a marked increase in the structural information derived from tandem mass spectra. Access and reuse of these queries is expected to enhance structural annotation in untargeted experiments, leading to more specific claims and advancing many applications in metabolomics.

Harwood, Thomas V. [USDOE Joint Genome Institute (↗

DOE OSTI2020

Rapid and Simultaneous Characterization of Drug Conjugation in Heavy and Light Chains of a Monoclonal Antibody Revealed by High-Resolution Ion Mobility Separations in SLIM

Antibody-drug conjugates (ADCs) have recently gained traction in the biomedical community due to their promise for human therapeutics and an alternative to chemotherapy for cancer. Crucial metrics for ADC efficacy, safety, and selectivity are their drug-antibody ratios (DARs). However, DAR characterization (i.e., determining the average number of conjugated drugs on the antibody) through analytical methods remains challenging due to the heterogeneity of drug conjugation as well as the numerous post-translational modifications possible in the monoclonal antibody. In this work, we report on the use of high-resolution ion mobility spectrometry separations in our structures for lossless ion manipulations coupled to mass spectrometry (SLIM IMS-MS) platform for the rapid and simultaneous characterization of the drug load profile (i.e., stoichiometric distribution of the number of conjugated drugs present on the mAb), determination of the weighted average DAR in both the heavy and light chains of a model antibody-drug conjugate, and calculation of the overall DAR of the ADC. After chemical reduction of the ADC and a subsequent 31.5 meter SLIM IMS separation the various drug bound antibody species could be well resolved for both chains. Furthermore, we also show that higher resolution separations were possible with SLIM IMS as compared to ones performed on a commercially available (~1 meter) drift tube IMS-MS platform. Interestingly, additional gas-phase substructures become evident with longer path length SLIM IMS, potentially providing insight related to how drug conjugation may alter the structural properties of a monoclonal antibody and thus affect its overall potency and/or selectivity for a given antigen. We expect high-resolution SLIM IMS separations will augment the existing toolbox for ADC characterization, particularly for structural confirmation to enable the rapid optimization of DAR for a given ADC and thus better understand its potential toxicity and potency.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2020

Methodology for the development of empirical models relating 13C NMR spectral features to fuel properties

Effective formulation of new gasoline or diesel fuels for internal combustion engines would benefit from the development of reliable models for predicting key fuel properties based on a set of molecular descriptors obtained from a single measurement. This is particularly relevant in the case of renewable fuels, where the available fuel sample quantity may be limited. In this work, we present a statistically-based methodology for building empirical models to predict multiple properties from one-dimensional 13C nuclear magnetic resonance (NMR) spectra measured on around 200 microliters of a liquid fuel. NMR spectra contain information about the molecular composition of a sample and the carbon types and molecular substructures therein. Our approach uses this information to build sparse, interpretable models, where the predicted properties are linked to specific molecular features. The approach takes into consideration the constrained nature of the features making up the one-dimensional NMR spectrum, which, after standardization, represent a relative fuel composition. We point to the limitations in interpretability that arise when building this type of empirical predictive model and suggest how these limitations may be diminished. Among the many properties important for maximizing engine performance and minimizing emissions, we build models that predict derived cetane number and distillation temperatures as these are of particular interest because of their links to fuel economy, drivability, and engine-out emissions. Results suggest that the properties of interest may be impacted by only a few of the 27 13C NMR regions represented in the data, pointing to new directions for further testing in the development of improved fuels.

predictive models, biofuels., NMR Analysis↗

DOE OSTI2020

In Silico Guidance for In Vitro Androgen and Glucocorticoid Receptor ToxCast Assays

Molecular initiating events (MIEs) are key events in adverse outcome pathways (AOPs) that link molecular chemistry to target biology. As they are based in chemistry, these interactions are excellent targets for computational chemistry approaches to in silico modelling. In this work, we aim to link ligand chemical structure to MIEs for androgen receptor (AR) and glucocorticoid receptor (GR) binding using ToxCast data. This has been done using an automated computational algorithm to perform maximal common substructure searches on chemical binders for each target from the ToxCast dataset. The models developed show a high level of accuracy, correctly assigning 87.20% of AR binders and 96.81% of GR binders in a 25% test set using holdout cross-validation. The 2D structural alerts developed can be used as in silico models to predict these MIEs, and as guidance for in vitro ToxCast assays to confirm hits. These models can target such experimental work, reducing the number of assays to be performed to gain required toxicological insight. Development of these models has also allowed some structural alerts to be identified as predictors for agonist or antagonist behavior at the receptor target. This work represents a first step in using computational methods to guide and target experimental approaches.

Allen, Timothy H.↗

DOE OSTI2021

Paramagnetic Molecular Semiconductors Combining Anisotropic Magnetic Ions with TCNQ Radical Anions

We report the synthesis, magnetic properties, and transport properties of paramagnetic metal complexes, [Co(DMF) 4 (TCNQ) 2 ](TCNQ) 2 (1), [La(DMF) 8 (TCNQ)](TCNQ) 5 (2), and [Nd(DMF) 7 (TCNQ)](TCNQ) 5 (3) (DMF = N,N-dimethylformamide, TCNQ = 7,7,8,8-tetracyanoquinodimethane). All three compounds contain fractionally charged TCNQ δ– anions (0 < δ < 1) and mononuclear complex cations in which the coordination environment of a metal center includes several DMF molecules and one or two terminally coordinated TCNQ δ– anions. The coordinated TCNQ δ– anions participate in π–π stacking interactions with noncoordinated TCNQ δ– anions, forming columnar substructures that provide efficient charge-transporting pathways. Furthermore, temperature-dependent conductivity measurements demonstrate that all three compounds exhibit semiconducting behavior.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2024

Advancing Heteroanionicity in Zintl Phases: Crystal Structures, Thermoelectric and Magnetic Properties of Two Quaternary Semiconducting Arsenide Oxides, Eu 8 Zn 2 As 6 O and Eu 14 Zn 5 As 12 O

Two novel quaternary oxyarsenides, Eu 8 Zn 2 As 6 O and Eu 14 Zn 5 As 12 O, were synthesized through metal flux reactions, and their crystal structures were established by single-crystal X-ray diffraction methods. Eu 8 Zn 2 As 6 O crystallizes in the orthorhombic space group Pbca, featuring polyanionic ribbons composed of corner-shared triangular [ZnAs 3 ] units, running along the [100] direction. The structure of Eu 14 Zn 5 As 12 O crystallizes in the monoclinic space group P2/m and its anionic substructure can be described as an infinite “ribbonlike” chain comprised of [ZnAs 3 ] trigonal-planar units, although the structural complexity here is greater and also amplified by disorder on multiple crystallographic positions. In both structures, the O 2– anion occupies an octahedral void with six neighboring Eu 2+ cations. Formal electron counting, electronic structure calculations, and transport properties reveal the charge-balanced semiconducting nature of these heteroanionic Zintl phases. High-temperature thermoelectric transport properties measurements on Eu 14 Zn 5 As 12 O reveal relatively high resistivity (ρ 500K = 8 Ω·cm) and Seebeck coefficient values (S 500K = 220 μV K –1 ), along with a low concentration and mobility of holes as the dominant charge-carriers (n 500K = 8.0 × 10 17 cm –3 , μ 500K = 6.4 cm 2 /V s). Magnetic studies indicate the presence of divalent Eu 2+ species in Eu 14 Zn 5 As 12 O and complex magnetic ordering, with two transitions observed at T 1 = 21.6 K and T 2 = 9 K.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2018

Synthesis, and Crystal and Electronic Structures, of the Titanium-Rich Bismuthides AE 3 Ti 8 Bi 10 ( AE = Sr, Ba, Eu)

Three isotypic compounds with the chemical formula AE 3 Ti 8 Bi 10 (AE = Sr, Ba, Eu) have been obtained via both high-temperature solid state and flux growth reactions. Their crystal structure, representing a new type (space group P6 3 /mmc, Pearson symbol hP 42 ), features an open framework composed of interlinked TiBi 5 square pyramids and TiBi 6 octahedra. The Ti–Bi substructure is penetrated by infinite columns of face-sharing AE 6 polyhedra centered by Bi atoms. First-principle calculations and physical property measurements indicate metallic behavior and absence of localized magnetic moments on the Ti atoms. Analysis of the chemical bonding reveals strong Ti–Bi and Ti–Ti bonds. Finally, the latter demonstrate classic two-center, as well as multicenter interactions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2019

In Situ Cs and H Exchange into Gaidonnayite and Proposed Mechanisms of Ion Diffusion

Here, the microporous mineral gaidonnayite Na 2 ZrSi 3 O 9 ·2H 2 O was studied to better understand its ion-exchange mechanisms, specifically for Cs + and H + ions. In situ Raman spectroscopy, in situ X-ray diffraction (XRD), simultaneous thermogravimetric analysis and differential scanning calorimetry (TGA/DSC), and in situ X-ray fluorescence were used to determine the exchange processes involved. The Raman spectra contain strong peaks that can be attributed to the vibrational modes for the 3MR symmetric stretch at 500 cm –1 , Si–O–Zr–O chain stretches at 938 cm –1 , and Si–O stretching in the 1000–1100 cm –1 range. The most prominent Raman shift during ion exchange is found near the 520 cm –1 peak, which corresponds to distortions of the 3MR substructure of gaidonnayite. In all instances of this study, the 3MR exhibited the highest amount of distortion during ion exchange, and the evolution of this distortion is compared to unit-cell changes as measured from XRD data and elemental changes via XRF. The correlations between the Raman, XRD, and XRF data show rapid deformation of the 3MR during the onset of H + ion exchange in the Na form of gaidonnayite. Even when unit-cell volume changes were small (<3 Å 3 ) as in the cases for Cs + into Na-gaidonnayite and Cs + into H-gaidonnayite, significant changes in the ≈520 cm –1 peak were measured. By comparing XRD data and Raman data, and verifying the cation uptake by XRF, we were able to identify and confirm conformational changes and distortions in the crystal structure before, during, and after Cs + and H + exchange. Finally, Cs exchange occurred the fastest and with the greatest capacity when starting in the H-form at room temperature, and at elevated temperatures when starting in the Na-form.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2018

Structural Characterization of Lignin from Maize ( Zea mays L.) Fibers: Evidence for Diferuloylputrescine Incorporated into the Lignin Polymer in Maize Kernels

The structure of the phenolic polymer in maize grain fibers, with 5.5% Klason lignin content, has been studied. For this, the milled wood lignin (MWL) and dioxane lignin (DL) preparations were isolated and analyzed. The data indicated that the lignin in maize fibers was syringyl rich, mostly involved in β-aryl ether, resinol, and phenylcoumaran substructures. 2D NMR and derivatization followed by reductive cleavage (DFRC) also revealed the occurrence of associated ferulates together with trace amounts of p-coumarates acylating the γ-OH of lignin side chains, predominantly on S-lignin units. More interesting was the occurrence of diferuloylputrescine, a ferulic acid amide, which was identified by 2D NMR and comparison with a synthesized standard, that was apparently incorporated into this lignin. A phenylcoumaran structure involving a diferuloylputrescine coupled through 8–5' linkages to another diferuloylputrescine (or to a ferulate or a guaiacyl lignin unit) was found, providing compelling evidence for its participation in radical coupling reactions. The occurrence of diferuloylputrescine in cell walls of maize kernels and other cereal grains appears to have been missed in previous works, perhaps due to the alkaline hydrolysis commonly used for composition studies.

2D NMR↗

DOE OSTI2018

Structural Characterization of Lignins from Willow Bark and Wood

Understanding the chemical structure of lignin in willow bark is an indispensable step to design how to separate its fiber bundles. The whole cell wall and enzyme lignin preparations sequentially isolated from ball-milled bark, inner bark, and wood were comparatively investigated by nuclear magnetic resonance (NMR) spectroscopy and three classical degradative methods, i.e., alkaline nitrobenzene oxidation, derivatization followed by reductive cleavage, and analytical thioacidolysis. All results demonstrated that the guaiacyl (G) units were predominant in the willow bark lignin over syringyl (S) and minor p-hydroxyphenyl (H) units. Moreover, the monomer yields and S/G ratio rose progressively from bark to inner bark and wood, indicating that lignin may be more condensed in bark than in other tissues. Additionally, major interunit linkage substructures (β-aryl ethers, phenylcoumarans, and resinols) together with cinnamyl alcohol end groups were relatively quantitated by two-dimensional NMR spectroscopy. Bark and inner bark were rich in pectins and proteins, which were present in large quantities and also in the enzyme lignin preparations.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2020

Chespa: Streamlining Expansive Chemical Space Evaluation of Molecular Sets

Thousands of chemical properties can be calculated for small molecules, which can be used to place the molecules within the context of a broader “chemical space.” These definitions vary based on compounds of interest and the goals for the given chemical space definition. Here, we introduce a customizable (i.e., modular) Python module, chespa, built to easily assess different chemical space definitions through cluster-ing of compounds in these spaces and visualize trends of these clusters. To demonstrate this, chespa currently streamlines prediction of vari-ous molecule descriptors (predicted chemical properties, molecular substructures, AI-based chemical space, and chemical class ontology) in order to test 6 different chemical space definitions. Furthermore, we investigated how these varying definitions trend with mass spectrometry (MS)-based observability, i.e., the ability of a molecule to be observed with MS (e.g., as a function of the molecule ionizability), using an example data set from the U.S. EPA's Non-Targeted Analysis Collaborative Trial (ENTACT), where blinded samples had been analyzed previously, providing 1,398 data points. Improved understanding of observability would offer many advantages in small molecule identifica-tion, such as (i) a priori selection of experimental conditions based on suspected sample composition, (ii) the ability to reduce the number of candidate structures during compound identification by removing those less likely to ionize, and, in turn, (iii) a reduced false discovery rate and increased confidence in identifications. Factors controlling observability are not fully understood, making prediction of this property non-trivial and a prime candidate for chemical space analysis. Chespa is available at github.com/pnnl/chespa.

Nunez, Jamie↗

DOE OSTI2025

RLMolLM: Reinforcement Learning-Enhanced Language Model Framework for Inverse Molecular Design

Inverse molecular design faces significant challenges due to vast chemical space and complex property requirements. While language models show promise for molecular generation, they struggle with validity, multi-property optimization, and structural constraints. This work presents RLMolLM, a reinforcement learning framework combining Proximal Policy Optimization (PPO) with genetic algorithms to address these limitations. Our approach optimizes multiple user-specified properties including quantitative estimates of drug-likeness (QED), synthetic accessibility (SA), and ADMET (absorption, distribution, metabolism, excretion, and toxicity) endpoints without requiring complete model retraining, while maintaining capability for scaffold-constrained generation where specific substructures must be preserved. We outperform state-of-the-art methods for molecular optimization, achieving best QED scores across GDB13, Moses, and Zinc datasets with up to 31% improvement over previous methods while maintaining excellent validity, uniqueness, and novelty metrics. For simultaneous multi-property optimization, our framework achieves substantial improvements in ADMET properties including 4.5-fold reduction in hERG toxicity and enhanced Caco-2 permeability compared to Moses dataset. Under structural constraints, the framework significantly improves molecular validity while preserving scaffolds and effectively optimizing properties. In conclusion, this versatile solution advances pharmaceutical and materials molecular design through effective integration of reinforcement learning and genetic algorithms with multi-property optimization and scaffold preservation.

Genetic algorithms↗