ABSTRACT
In materials science, accurately computing properties like viscosity, melting point, and glass transition temperatures solely through physics-based models is challenging. Data-driven machine learning (ML) also poses challenges in constructing ML models, especially in the material science domain where data is limited. To address this, we integrate physics-informed descriptors from molecular dynamics (MD) simulations to enhance the accuracy and interpretability of ML models. Our current study focuses on accurately predicting viscosity in liquid systems using MD descriptors. In this work, we curated a comprehensive dataset of over 4000 small organic molecules' viscosities from scientific literature, publications, and online databases. This dataset enabled us to develop quantitative structure-property relationships (QSPR) consisting of descriptor-based and graph neural network models to predict temperature-dependent viscosities for a wide range of viscosities. The QSPR models reveal that including MD descriptors improves the prediction of experimental viscosities, particularly at the small data set scale of fewer than a thousand data points. Furthermore, feature importance tools reveal that intermolecular interactions captured by MD descriptors are most important for viscosity predictions. Finally, the QSPR models can accurately capture the inverse relationship between viscosity and temperature for six battery-relevant solvents, some of which were not included in the original data set. Our research highlights the effectiveness of incorporating MD descriptors into QSPR models, which leads to improved accuracy for properties that are difficult to predict when using physics-based models alone or when limited data is available.
ABSTRACT
The blood-brain barrier (BBB) plays a critical role in preventing harmful endogenous and exogenous substances from penetrating the brain. Optimal brain penetration of small-molecule central nervous system (CNS) drugs is characterized by a high unbound brain/plasma ratio (Kp,uu). While various medicinal chemistry strategies and in silico models have been reported to improve BBB penetration, they have limited application in predicting Kp,uu directly. We describe a physics-based computational approach, a quantum mechanics (QM)-based energy of solvation (E-sol), to predict Kp,uu. Prospective application of this method in internal CNS drug discovery programs highlights the utility and accuracy of this new method, which showed a categorical accuracy of 79% and an R2 of 0.61 from a linear regression model.
Subject(s)
Blood-Brain Barrier , Brain , Biological Transport/physiology , Central Nervous System Agents , Computer SimulationABSTRACT
Epik version 7 is a software program that uses machine learning for predicting the pKa values and protonation state distribution of complex, druglike molecules. Using an ensemble of atomic graph convolutional neural networks (GCNNs) trained on over 42,000 pKa values across broad chemical space from both experimental and computed origins, the model predicts pKa values with 0.42 and 0.72 pKa unit median absolute and root mean square errors, respectively, across seven test sets. Epik version 7 also generates protonation states and recovers 95% of the most populated protonation states compared to previous versions. Requiring on average only 47 ms per ligand, Epik version 7 is rapid and accurate enough to evaluate protonation states for crucial molecules and prepare ultra-large libraries of compounds to explore vast regions of chemical space. The simplicity and time required for the training allow for the generation of highly accurate models customized to a program's specific chemistry.
ABSTRACT
Wnt family proteins and ß-catenin are critical for the regulation of many developmental and oncogenic processes. Wnts are secreted protein ligands which signal using a canonical pathway, and involve the transcriptional co-activator ß-catenin or non-canonical pathways that are independent of ß-catenin. Bone metastasis is unfortunately a common occurrence in prostate cancer and can be conceptualized as a series of related steps or processes, most of which are regulated by Wnt ligands and/or ß-catenin. At the primary tumor site, cancer cells often take on mesenchymal properties, termed epithelial mesenchymal transition (EMT), which are regulated in part by the Wnt receptor FZD4. Then, Wnt signaling, especially Wnt5A, is of importance as the cells circulate in the blood stream. Upon arriving in the bones, cancer cells migrate and take on stem-like or tumorigenic properties, as aided through Wnt or ß-catenin signaling involving CHD11, CD24, and Wnt5A. Additionally, cancer cells can become dormant and evade therapy, in part due to regulation by Wnt5A. In the bones, E-selectin can aid in the reversal of EMT, a process termed mesenchymal epithelial transition (MET), as a part of metastatic tumorigenesis. Once bone tumors are established, Wnt/ß-catenin signaling is involved in the suppression of osteoblast function largely through DKK1.
ABSTRACT
The lone star tick, Amblyomma americanum Linnaeus, is ubiquitously present in the southeastern United States and will readily parasitize humans and companion animals. Fipronil is the active ingredient in several topically applied products used to manage ticks and fleas on companion animals. Reducing ectoparasite infestations on companion animals decreases the risk that pathogens vectored by these pests are transmitted to these animals and their owners. However, dependence on acaricides can promote the development of resistance. In this study, the Food and Agriculture Organization of the United Nations larval packet test was used to determine the lethal concentration (LC) values and discriminating concentration (DC) for fipronil in the lone star tick. The DC was calculated as 0.02080%. The low magnitude of the DC value suggests that fipronil is an effective active ingredient for A. americanum management. With the LC and DC values determined, emergent resistance can be monitored, potentially allowing for intervention should tolerance develop in A. americanum populations that are in frequent contact with humans and their companion animals.
Subject(s)
Acaricides , Amblyomma , Pyrazoles , Tick Control , Amblyomma/growth & development , Animals , Drug Resistance , Larva/growth & developmentABSTRACT
The aim of this paper is to provide new theoretical and computational understanding on two loss regularizations employed in deep learning, known as local entropy and heat regularization. For both regularized losses, we introduce variational characterizations that naturally suggest a two-step scheme for their optimization, based on the iterative shift of a probability density and the calculation of a best Gaussian approximation in Kullback-Leibler divergence. Disregarding approximation error in these two steps, the variational characterizations allow us to show a simple monotonicity result for training error along optimization iterates. The two-step optimization schemes for local entropy and heat regularized loss differ only over which argument of the Kullback-Leibler divergence is used to find the best Gaussian approximation. Local entropy corresponds to minimizing over the second argument, and the solution is given by moment matching. This allows replacing traditional backpropagation calculation of gradients by sampling algorithms, opening an avenue for gradient-free, parallelizable training of neural networks. However, our presentation also acknowledges the potential increase in computational cost of naive optimization of regularized costs, thus giving a less optimistic view than existing works of the gains facilitated by loss regularization.
ABSTRACT
Amanita is a diverse and cosmopolitan genus of ectomycorrhizal fungi. We describe Amanita nouhrae sp. nov., a new hypogeous ('truffle-like') species associated with Nothofagus antarctica in northern Patagonia. This constitutes the first report of a sequestrate Amanita from the Americas. Thick-walled basidiospores ornamented on the interior spore wall ('crassospores') were observed consistently in A. nouhrae and its sister epigeous taxon Amanita morenoi, a rarely collected but apparently common species from northern Patagonia that has sometimes been misidentified as the Australian taxon Amanita umbrinella. Nuclear 18S and 28S ribosomal DNA and mitochondrial 16S and 26S DNA placed these two species in a southern temperate clade within subgenus Amanita, together with other South American and Australian species. Based on a dated genus-level phylogeny, we estimate that the southern temperate clade may have originated near the Eocene/Oligocene boundary (ca. 35 Ma ± 10 Ma). This date suggests a broadly distributed ancestor in the Southern Hemisphere, which probably diversified as a result of continental drift, as well as the initiation of the Antarctic glaciation. By comparison, we show that this clade follows an exceptional biogeographic pattern within a genus otherwise seemingly dominated by Northern Hemisphere dispersal.