Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “prediction”

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.

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

At least 1,081 records · Page 60

A New Model for Predicting Acute Mucosal Toxicity in Head-and-Neck Cancer Patients Undergoing Radiotherapy With Altered Schedules

Purpose: One of the worst radiation-induced acute effects in treating head-and-neck (HN) cancer is grade 3 or higher acute (oral and pharyngeal) mucosal toxicity (AMT), caused by the killing/depletion of mucosa cells. Here we aim to testing a predictive model of the AMT in HN cancer patients receiving different radiotherapy schedules. Methods and Materials: Various radiotherapeutic schedules have been reviewed and classified as tolerable or intolerable based on AMT severity. A modified normal tissue complication probability (NTCP) model has been investigated to describe AMT data in radiotherapy regimens, both conventional and altered in dose and overall treatment time (OTT). We tested the hypothesis that such a model could also be applied to identify intolerable treatment and to predict AMT. This AMT NTCP model has been compared with other published predictive models to identify schedules that are either tolerable or intolerable. The area under the curve (AUC) was calculated for all models, assuming treatment tolerance as the gold standard. The correlation between AMT and the predicted toxicity rate was assessed by a Pearson correlation test. Results: The AMT NTCP model was able to distinguish between acceptable and intolerable schedules among the data available for the study (AUC = 0.84, 95% confidence interval = 0.75-0.92). In the equivalent dose at 2 Gy/fraction (EQD2) vs OTT space, the proposed model shows a trend similar to that of models proposed by other authors, but was superior in detecting some intolerable schedules. Moreover, it was able to predict the incidence of {>=}G3 AMT. Conclusion: The proposed model is able to predict {>=}G3 AMT after HN cancer radiotherapy, and could be useful for designing altered/hypofractionated schedules to reduce the incidence of AMT.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Pitfalls in Prediction Modeling for Normal Tissue Toxicity in Radiation Therapy: An Illustration With the Individual Radiation Sensitivity and Mammary Carcinoma Risk Factor Investigation Cohorts

Purpose: To identify the main causes underlying the failure of prediction models for radiation therapy toxicity to replicate. Methods and Materials: Data were used from two German cohorts, Individual Radiation Sensitivity (ISE) (n=418) and Mammary Carcinoma Risk Factor Investigation (MARIE) (n=409), of breast cancer patients with similar characteristics and radiation therapy treatments. The toxicity endpoint chosen was telangiectasia. The LASSO (least absolute shrinkage and selection operator) logistic regression method was used to build a predictive model for a dichotomized endpoint (Radiation Therapy Oncology Group/European Organization for the Research and Treatment of Cancer score 0, 1, or ≥2). Internal areas under the receiver operating characteristic curve (inAUCs) were calculated by a naïve approach whereby the training data (ISE) were also used for calculating the AUC. Cross-validation was also applied to calculate the AUC within the same cohort, a second type of inAUC. Internal AUCs from cross-validation were calculated within ISE and MARIE separately. Models trained on one dataset (ISE) were applied to a test dataset (MARIE) and AUCs calculated (exAUCs). Results: Internal AUCs from the naïve approach were generally larger than inAUCs from cross-validation owing to overfitting the training data. Internal AUCs from cross-validation were also generally larger than the exAUCs, reflecting heterogeneity in the predictors between cohorts. The best models with largest inAUCs from cross-validation within both cohorts had a number of common predictors: hypertension, normalized total boost, and presence of estrogen receptors. Surprisingly, the effect (coefficient in the prediction model) of hypertension on telangiectasia incidence was positive in ISE and negative in MARIE. Other predictors were also not common between the 2 cohorts, illustrating that overcoming overfitting does not solve the problem of replication failure of prediction models completely. Conclusions: Overfitting and cohort heterogeneity are the 2 main causes of replication failure of prediction models across cohorts. Cross-validation and similar techniques (eg, bootstrapping) cope with overfitting, but the development of validated predictive models for radiation therapy toxicity requires strategies that deal with cohort heterogeneity.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Predicting Patient-specific Dosimetric Benefits of Proton Therapy for Skull-base Tumors Using a Geometric Knowledge-based Method

Purpose: To predict the organ at risk (OAR) dose levels achievable with proton beam therapy (PBT), solely based on the geometric arrangement of the target volume in relation to the OARs. A comparison with an alternative therapy yields a prediction of the patient-specific benefits offered by PBT. This could enable physicians at hospitals without proton capabilities to make a better-informed referral decision or aid patient selection in model-based clinical trials. Methods and Materials: Skull-base tumors were chosen to test the method, owing to their geometric complexity and multitude of nearby OARs. By exploiting the correlations between the dose and distance-to-target in existing PBT plans, the models were independently trained for 6 types of OARs: brainstem, cochlea, optic chiasm, optic nerve, parotid gland, and spinal cord. Once trained, the models could estimate the feasible dose–volume histogram and generalized equivalent uniform dose (gEUD) for OAR structures of new patients. The models were trained using 20 patients and validated using an additional 21 patients. Validation was achieved by comparing the predicted gEUD to that of the actual PBT plan. Results: The predicted and planned gEUD were in good agreement. Considering all OARs, the prediction error was +1.4 ± 5.1 Gy (mean ± standard deviation), and Pearson's correlation coefficient was 93%. By comparing with an intensity modulated photon treatment plan, the model could classify whether an OAR structure would experience a gain, with a sensitivity of 93% (95% confidence interval: 87%-97%) and specificity of 63% (95% confidence interval: 38%-84%). Conclusions: We trained and validated models that could quickly and accurately predict the patient-specific benefits of PBT for skull-base tumors. Similar models could be developed for other tumor sites. Such models will be useful when an estimation of the feasible benefits of PBT is desired but the experience and/or resources required for treatment planning are unavailable.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Neural operator prediction of linear instability waves in high-speed boundary layers

Here, we investigate if neural operators can predict the linear evolution of instability waves in high-speed boundary layers. To this end, we extend the design of the DeepOnet to ensure accurate and robust predictions, and also to perform data assimilation. In particular, we train DeepONet to take as inputs an upstream disturbance and a downstream location of interest, and to provide as output the perturbation field downstream in the boundary layer. DeepONet thus approximates the linearized and parabolized Navier-Stokes operator for this flow. For successful application to the high-speed boundary layer problem, we add sample weighting and Fourier input features to the regular DeepONet formulation. Once trained, the DeepOnet can perform fast and accurate predictions of the downstream disturbances within the range of training frequencies (inside the distribution). In addition, we show that DeepONet can solve the inverse problem, where downstream wall measurements are adopted as input, and a trained network can predict the upstream disturbances that led to these observations. This capability, along with the forward predictions, allows us to perform a full data assimilation cycle efficiently: starting from wall-pressure data, we predict the upstream disturbance using the inverse DeepONet and its evolution using the forward DeepONet. Finally, we introduce three new metrics to benchmark the training, evaluation and break-even cost of neural operators.

42 ENGINEERING↗

Maximum likelihood Bayesian model averaging and its predictive analysis for groundwater reactive transport models

While Bayesian model averaging (BMA) has been widely used in groundwater modeling, it is infrequently applied to groundwater reactive transport modeling because of multiple sources of uncertainty in the coupled hydrogeochemical processes and because of the long execution time of each model run. To resolve these problems, this study analyzed different levels of uncertainty in a hierarchical way, and used the maximum likelihood version of BMA, i.e., MLBMA, to improve the computational efficiency. Our study demonstrates the applicability of MLBMA to groundwater reactive transport modeling in a synthetic case in which twenty-seven reactive transport models were designed to predict the reactive transport of hexavalent uranium (U(VI)) based on observations at a former uranium mill site near Naturita, CO. Moreover, these reactive transport models contain three uncertain model components, i.e., parameterization of hydraulic conductivity, configuration of model boundary, and surface complexation reactions that simulate U(VI) adsorption. These uncertain model components were aggregated into the alternative models by integrating a hierarchical structure into MLBMA. The modeling results of the individual models and MLBMA were analyzed to investigate their predictive performance. The predictive logscore results show that MLBMA generally outperforms the best model, suggesting that using MLBMA is a sound strategy to achieve more robust model predictions relative to a single model. MLBMA works best when the alternative models are structurally distinct and have diverse model predictions. When correlation in model structure exists, two strategies were used to improve predictive performance by retaining structurally distinct models or assigning smaller prior model probabilities to correlated models. Since the synthetic models were designed using data from the Naturita site, the results of this study are expected to provide guidance for real-world modeling. Finally, limitations of applying MLBMA to the synthetic study and future real-world modeling are discussed.

58 GEOSCIENCES↗

Utilizing machine learning to predict tensile ductility and yield strength of CoNiV-based multi-principal elements alloys

This study explores the use of machine learning (ML) as a computational tool to accelerate the design of multi-principal element alloys (MPEAs) with improved tensile elongation. An ML model was trained using available experimental data from the literature along with theoretically derived features to predict yield strength (YS) and ductility. A subset of ML-predicted compositions—CoNiVFe, CoNiVTi, CoNiVTiFe, and CoCrNiVTi—was synthesized and evaluated through tensile testing. The ML model underpredicted YS by approximately 20–30 % and overpredicted ductility by 60–70 % for Ti-containing alloys. Microstructural analysis revealed that Ti segregation at interdendritic regions contributed to early fracture, leading to discrepancies in ductility predictions. Ti segregation at these regions likely drives the increased YS due to segregation strengthening. In contrast, the CoNiVFe alloy showed good agreement with both experimental YS and elongation, with prediction errors of ∼10.2 % and ∼20.7 %, respectively. Microstructural characterization revealed minimal segregation in this alloy, suggesting that the ML model can reliably predict the properties of alloys with little to no segregation. These findings highlight the capability of ML in predicting YS with good accuracy but underscore its limitations in capturing defect-driven failure mechanisms such as segregation-induced embrittlement.

36 MATERIALS SCIENCE↗

Method of predicting the mean lung dose based on a patient's anatomy and dose-volume histograms

The aim of this study was to propose a method to predict the minimum achievable mean lung dose (MLD) and corresponding dosimetric parameters for organs-at-risk (OAR) based on individual patient anatomy. For each patient, the dose for 36 equidistant individual multileaf collimator shaped fields in the treatment planning system (TPS) was calculated. Based on these dose matrices, the MLD for each patient was predicted by the homemade DosePredictor software in which the solution of linear equations was implemented. The software prediction results were validated based on 3D conformal radiotherapy (3D-CRT) and volumetric modulated arc therapy (VMAT) plans previously prepared for 16 patients with stage III non–small-cell lung cancer (NSCLC). For each patient, dosimetric parameters derived from plans and the results calculated by DosePredictor were compared. The MLD, the maximum dose to the spinal cord (D{sub max} {sub cord}) and the mean esophageal dose (MED) were analyzed. There was a strong correlation between the MLD calculated by the DosePredictor and those obtained in treatment plans regardless of the technique used. The correlation coefficient was 0.96 for both 3D-CRT and VMAT techniques. In a similar manner, MED correlations of 0.98 and 0.96 were obtained for 3D-CRT and VMAT plans, respectively. The maximum dose to the spinal cord was not predicted very well. The correlation coefficient was 0.30 and 0.61 for 3D-CRT and VMAT, respectively. The presented method allows us to predict the minimum MLD and corresponding dosimetric parameters to OARs without the necessity of plan preparation. The method can serve as a guide during the treatment planning process, for example, as initial constraints in VMAT optimization. It allows the probability of lung pneumonitis to be predicted.

61 RADIATION PROTECTION AND DOSIMETRY↗

Chemical composition based machine learning model to predict defect formation in additive manufacturing

With a goal of exploiting additive manufacturing to improve the manufacturing of existing reactor materials, we developed a chemical composition-based machine learning model to predict the printability of any given alloy in laser powder bed fusion (L-PBF) using experimental data from peer-reviewed literature. We defined printability as the ability to avoid defects like cracking, balling, porosity, and lack of fusion, that are caused by thermal stresses (during solidification or liquation), molten pool disintegration into disconnected small beads or lack of heat input respectively. Our models predict the tendency of balling defect formation and porosity percentage for a given composition, under a given set of processing conditions. To predict the likelihood of balling defect, three models: a random forest classifier, a gradient boost regressor and a neural network were trained on a dataset containing both traditional alloys and high entropy alloys. The neural network model showed the highest accuracy of 92.3 % in predicting the balling defect formation. A random forest regressor, gradient boost regressor and neural network were trained and tested on a dataset of various alloys to predict porosity. The random forest regressor showed the best predictions with an R 2 score of 0.97. The models also revealed the relative importance of the input descriptors on defect-formation tendency. Of particular significance was the identification of carbon as an important element in determining the occurrence of balling and percent porosity in alloys like steel, as well as being moderately important to the percentage porosity in other alloys as well as steel. Manganese was also identified as a key descriptor for the percentage of porosity in steel and other alloys. Manganese’s low thermal conductivity and consistent presence in the dataset is the likely cause for its contribution. Carbon’s role is attributable to its relatively high specific heat and high melting temperature. In conclusion, our model serves as a swift, chemistry-based tool to design experiments and find modified compositions better suited for additive manufacturing.

36 MATERIALS SCIENCE↗

Improved departure from nucleate boiling prediction in rod bundles using a physics-informed machine learning-aided framework

The critical heat flux (CHF) corresponding to the departure from nucleate boiling (DNB) crisis is a regulatory limit for the licensing of pressurized water reactors (PWRs) worldwide. Despite the abundance of predictive tools available to the reactor thermal-hydraulics community, the path for an accurate CHF model remains elusive. This work approaches the prediction of DNB through a physics-informed machine learning-aided framework (PIMLAF) with the objective of achieving superior predictive capabilities for a rod bundle. In view of the limitations in existing macro-scale physics-driven tools, an improved mechanistic model is first proposed, leveraging key concepts in the liquid sublayer dryout and bubble crowding mechanisms. Furthermore, the proposed mechanistic model is able to predict DNB in different heater geometries for a broad range of flow conditions without the need for recalibration. This model is then incorporated as the physics-informed component of the hybrid framework PIMLAF, which takes advantage of established understanding in the field (i.e., domain knowledge [DK]) and uses machine learning (ML) to capture undiscovered information from the mismatch between the actual and DK-predicted output. Two bundle-related case studies using the PWR subchannel and bundle tests (PSBT) database are carried out to illustrate the PIMLAF’s improved performance over traditional approaches for both interpolation and extrapolation purposes. In light of the PIMLAF’s promising potential to reduce prediction error, reactor vendors are encouraged to leverage their in-house experimental efforts and apply the hybrid framework to potentially achieve margin reductions in the minimum DNB ratio (MDNBR) for the designs of interest.

42 ENGINEERING↗

Multiphysics simulation of TRISO fuel compacts and the effects of homogenization on silver release predictions

This work studies the impact of explicit and homogenized modeling approaches on the multiphysics simulation of TRistructural ISOtropic (TRISO) fuel compacts in prismatic High Temperature Gas Reactors (HTGRs) and the silver release predictions. TRISO fuel compacts exhibit complex double heterogeneity that significantly affects heat conduction, neutron transport, and fission products release. In this work, we use Cardinal, a multiphysics tool based on the Multiphysics Object-Oriented Simulation Environment (MOOSE) framework, to couple neutron transport and heat conduction. OpenMC is used for neutron transport and the MOOSE heat transfer module is used for the heat conduction. Then, we use BISON for the silver release predictions. Two different modeling approaches—explicit modeling of individual TRISO particles and homogenized representation using effective thermal properties—are compared at high TRISO packing fractions (20% and 40%) across varying power densities. Results demonstrate that homogenization significantly under-predicts peak temperatures, for the case of high power/TRISO, there is a difference of 83.73 K in the maximum temperature. Additionally, homogenization underestimates the silver release predictions compared to explicit modeling. Specifically, the temperature and power density differences lead to significant differences in silver release predictions. This work demonstrates the importance of accurately modeling heterogeneity of the TRISO particles to reliably predict fission product release and assess reactor safety margins.

11 - NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

A novel peridynamics-based approach to predict pharmaceutical tablet robustness

The pharmaceutical drug product development process can be greatly accelerated through the use of modeling and simulation techniques to predict the manufacturability and performance of a given formulation. The anticipation and possible mitigation of tablet damage due to manufacturing stresses represents a specific area of interest in the pharmaceutical industry for predicting formulation and tableting performance. While the finite element method (FEM) has been extensively used for predicting the mechanical behavior of powder material in the compaction processes, a shortcoming of the approach is the inherent difficulty to predict discontinuities (e.g., damage or cracking) within a tablet as FEM is a continuum-based approach. In this work, we propose a novel method utilizing peridynamics (PD), a numerical method that can capture discontinuities such as tablet fracture, to predict the evolution of damage and breakage in pharmaceutical tablets. The approach links (1) the finite element method – to elucidate the behavior of powders during die compaction – with (2) the peridynamics modeling technique – to model the discontinuous nature of damage and predict tablet breakage during the critical stages of unloading and ejection from the compression die. This short communication presents a proof of concept including a workflow to calibrate the linked FEM-PD simulation models. Further, it demonstrates promising results from a preliminary experimental validation of the approach. Following further development, this approach could be used to guide the optimization of compression processes through targeted changes to formulation material properties, compression process conditions, and/or tooling geometries to deliver improved process efficiency and tablet robustness.

36 MATERIALS SCIENCE↗

A hybrid approach to advancing quantitative prediction of tissue distribution of basic drugs in human

A general toxicity of basic drugs is related to phospholipidosis in tissues. Therefore, it is essential to predict the tissue distribution of basic drugs to facilitate an initial estimate of that toxicity. The objective of the present study was to further assess the original prediction method that consisted of using the binding to red blood cells measured in vitro for the unbound drug (RBCu) as a surrogate for tissue distribution, by correlating it to unbound tissue:plasma partition coefficients (Kpu) of several tissues, and finally to predict volume of distribution at steady-state (V{sub ss}) in humans under in vivo conditions. This correlation method demonstrated inaccurate predictions of V{sub ss} for particular basic drugs that did not follow the original correlation principle. Therefore, the novelty of this study is to provide clarity on the actual hypotheses to identify i) the impact of pharmacological mode of action on the generic correlation of RBCu-Kpu, ii) additional mechanisms of tissue distribution for the outlier drugs, iii) molecular features and properties that differentiate compounds as outliers in the original correlation analysis in order to facilitate its applicability domain alongside the properties already used so far, and finally iv) to present a novel and refined correlation method that is superior to what has been previously published for the prediction of human V{sub ss} of basic drugs. Applying a refined correlation method after identifying outliers would facilitate the prediction of more accurate distribution parameters as key inputs used in physiologically based pharmacokinetic (PBPK) and phospholipidosis models.

60 APPLIED LIFE SCIENCES↗

Pharmacokinetic modeling: Prediction and evaluation of route dependent dosimetry of bisphenol A in monkeys with extrapolation to humans

A physiologically based pharmacokinetic (PBPK) model was developed for bisphenol A (BPA) in adult rhesus monkeys using intravenous (iv) and oral bolus doses of 100 {mu}g d6-BPA/kg (). This calibrated PBPK adult monkey model for BPA was then evaluated against published monkey kinetic studies with BPA. Using two versions of the adult monkey model based on monkey BPA kinetic data from and , the aglycone BPA pharmacokinetics were simulated for human oral ingestion of 5 mg d16-BPA per person (Voelkel et al., 2002). Voelkel et al. were unable to detect the aglycone BPA in plasma, but were able to detect BPA metabolites. These human model predictions of the aglycone BPA in plasma were then compared to previously published PBPK model predictions obtained by simulating the Voelkel et al. kinetic study. Our BPA human model, using two parameter sets reflecting two adult monkey studies, both predicted lower aglycone levels in human serum than the previous human BPA PBPK model predictions. BPA was metabolized at all ages of monkey (PND 5 to adult) by the gut wall and liver. However, the hepatic metabolism of BPA and systemic clearance of its phase II metabolites appear to be slower in younger monkeys than adults. The use of the current non-human primate BPA model parameters provides more confidence in predicting the aglycone BPA in serum levels in humans after oral ingestion of BPA. -- Highlights: Black-Right-Pointing-Pointer A bisphenol A (BPA) PBPK model for the infant and adult monkey was constructed. Black-Right-Pointing-Pointer The hepatic metabolic rate of BPA increased with age of the monkey. Black-Right-Pointing-Pointer The systemic clearance rate of metabolites increased with age of the monkey. Black-Right-Pointing-Pointer Gut wall metabolism of orally administered BPA was substantial across all ages of monkeys. Black-Right-Pointing-Pointer Aglycone BPA plasma concentrations were predicted in humans orally given oral doses of deuterated BPA.

60 APPLIED LIFE SCIENCES↗

Predicting changes in cardiac myocyte contractility during early drug discovery with in vitro assays

Cardiovascular-related adverse drug effects are a major concern for the pharmaceutical industry. Activity of an investigational drug at the L-type calcium channel could manifest in a number of ways, including changes in cardiac contractility. The aim of this study was to define which of the two assay technologies – radioligand-binding or automated electrophysiology – was most predictive of contractility effects in an in vitro myocyte contractility assay. The activity of reference and proprietary compounds at the L-type calcium channel was measured by radioligand-binding assays, conventional patch-clamp, automated electrophysiology, and by measurement of contractility in canine isolated cardiac myocytes. Activity in the radioligand-binding assay at the L-type Ca channel phenylalkylamine binding site was most predictive of an inotropic effect in the canine cardiac myocyte assay. The sensitivity was 73%, specificity 83% and predictivity 78%. The radioligand-binding assay may be run at a single test concentration and potency estimated. The least predictive assay was automated electrophysiology which showed a significant bias when compared with other assay formats. Given the importance of the L-type calcium channel, not just in cardiac function, but also in other organ systems, a screening strategy emerges whereby single concentration ligand-binding can be performed early in the discovery process with sufficient predictivity, throughput and turnaround time to influence chemical design and address a significant safety-related liability, at relatively low cost. - Highlights: • The L-type calcium channel is a significant safety liability during drug discovery. • Radioligand-binding to the L-type calcium channel can be measured in vitro. • The assay can be run at a single test concentration as part of a screening cascade. • This measurement is highly predictive of changes in cardiac myocyte contractility.

60 APPLIED LIFE SCIENCES↗

The value of integrating pre-clinical data to predict nausea and vomiting risk in humans as illustrated by AZD3514, a novel androgen receptor modulator

Nausea and vomiting are components of a complex mechanism that signals food avoidance and protection of the body against the absorption of ingested toxins. This response can also be triggered by pharmaceuticals. Predicting clinical nausea and vomiting liability for pharmaceutical agents based on pre-clinical data can be problematic as no single animal model is a universal predictor. Moreover, efforts to improve models are hampered by the lack of translational animal and human data in the public domain. AZD3514 is a novel, orally-administered compound that inhibits androgen receptor signaling and down-regulates androgen receptor expression. Here we have explored the utility of integrating data from several pre-clinical models to predict nausea and vomiting in the clinic. Single and repeat doses of AZD3514 resulted in emesis, salivation and gastrointestinal disturbances in the dog, and inhibited gastric emptying in rats after a single dose. AZD3514, at clinically relevant exposures, induced dose-responsive “pica” behaviour in rats after single and multiple daily doses, and induced retching and vomiting behaviour in ferrets after a single dose. We compare these data with the clinical manifestation of nausea and vomiting encountered in patients with castration-resistant prostate cancer receiving AZD3514. Our data reveal a striking relationship between the pre-clinical observations described and the experience of nausea and vomiting in the clinic. In conclusion, the emetic nature of AZD3514 was predicted across a range of pre-clinical models, and the approach presented provides a valuable framework for predicition of clinical nausea and vomiting. - Highlights: • Integrated pre-clinical data can be used to predict clinical nausea and vomiting. • Data integrated from standard toxicology studies is sufficient to make a prediction. • The use of the nausea algorithm developed by Parkinson (2012) aids the prediction. • Additional pre-clinical studies can be used to confirm and quantify the risk.

60 APPLIED LIFE SCIENCES↗

Machine learning prediction of glass transition temperature of conjugated polymers from chemical structure

Predicting the glass transition temperature (T g ) is of critical importance as it governs the thermomechanical performance of conjugated polymers (CPs). Here, we report a predictive modeling framework to predict T g of CPs through the integration of machine learning (ML), molecular dynamics (MD) simulations, and experiments. With 154 T g data collected, an ML model is developed by taking simplified “geometry” of six chemical building blocks as molecular features, where side-chain fraction, isolated rings, fused rings, and bridged rings features are identified as the dominant ones for T g . MD simulations further unravel the fundamental roles of those chemical building blocks in dynamical heterogeneity and local mobility of CPs at a molecular level. The developed ML model is demonstrated for its capability of predicting T g of several new high-performance solar cell materials to a good approximation. The established predictive framework facilitates the design and prediction of T g of complex CPs, paving the way for addressing device stability issues that have hampered the field from developing stable organic electronics.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Importance of Engineered and Learned Molecular Representations in Predicting Organic Reactivity, Selectivity, and Chemical Properties

Machine-readable chemical structure representations are foundational in all attempts to harness machine learning for the prediction of reactivities, selectivities, and chemical properties directly from molecular structure. The featurization of discrete chemical structures into a continuous vector space is a critical phase undertaken before model selection, and the development of new ways to quantitatively encode molecules is an active area of research. Here, we highlight the application and suitability of different representations, from expert-guided “engineered” descriptors to automatically “learned” features, in different prediction tasks relevant to organic and organometallic chemistry, where differing amounts of training data are available. These tasks include statistical models of stereo- and enantioselectivity, thermochemistry, and kinetics developed using experimental and quantum chemical data. The use of expert-guided molecular descriptors provides an opportunity to incorporate chemical knowledge, domain expertise, and physical constraints into statistical modeling. In applications to stereoselective organic and organometallic catalysis, where data sets may be relatively small and 3D-geometries and conformations play an important role, mechanistically informed features can be used successfully to obtain predictive statistical models that are also chemically interpretable. We provide an overview of several recent applications of this approach to obtain quantitative models for reactivity and selectivity, where topological descriptors, quantum mechanical calculations of electronic and steric properties, along with conformational ensembles, all feature as essential ingredients of the molecular representations used. Alternatively, more flexible, general-purpose molecular representations such as attributed molecular graphs can be used with machine learning approaches to learn the complex relationship between a structure and prediction target. This approach has the potential to out-perform more traditional representation methods such as “hand-crafted” molecular descriptors, particularly as data set sizes grow. One area where this is particularly relevant is in the use of large sets of quantum mechanical data to train quantitative structure–property relationships. A general approach toward curating useful data sets and training highly accurate graph neural network models is discussed in the context of organic bond dissociation enthalpies, where this strategy outperforms regression using precomputed descriptors. Finally, we describe how graph neural network predictions can be incorporated into mechanistically informed statistical models of chemical reactivity and selectivity. Once trained, this approach avoids the expensive computational overhead associated with quantum mechanical calculations, while maintaining chemical interpretability. We illustrate examples for which fast predictions of bond dissociation enthalpy and of the identities of radicals formed through cleavage of a molecule’s weakest bond are used in simple physical models of site-selectivity and reactivity.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

CFM-ID 4.0: More Accurate ESI MS/MS Spectral Prediction and Compound Identification

In the field of metabolomics, mass spectrometry (MS) is the method most commonly used for identifying and annotating metabolites. As this typically involves matching a given MS spectrum against an experimentally acquired reference spectral library, this approach is limited by the coverage and size of such libraries (which typically number in the thousands). These experimental libraries can be greatly extended by predicting the MS spectra of known chemical structures (which number in the millions) to create computational reference spectral libraries. To facilitate the generation of predicted spectral reference libraries we developed CFM-ID, a computer program that can accurately predict ESI-MS/MS spectrum for a given compound structure. CFM-ID is one of the best-performing methods for compound-to-mass-spectrum prediction, and also one of the top tools for in silico mass-spectrum-to-compound identification. This work improves CFM-ID’s ability to predict ESI-MS/MS spectra from compounds by: (1) learning parameters from features based on the molecular topology, (2) adding a new approach to ring cleavage that models such cleavage as a sequence of simple chemical bond dissociations and (3) expanding its hand-written rule-based predictor to cover more chemical classes, including acylcarnitines, acylcholines, flavonols, flavones, flavanones, and flavonoid glycosides. We demonstrate that this new version of CFM-ID (version 4.0) is significantly more accurate than previous CFM-ID versions, in terms of both EI-MS/MS spectral prediction and compound identification. CFM-ID 4.0 is available at http://cfmid4.wishartlab.com/ as a webservice and docker images can be downloaded at https://hub.docker.com/r/wishartlab/cfmid

Wang, Fei↗