Disentangling Socioeconomic Status and Race in Infant Brain, Birth Weight, and Gestational Age at Birth: A Neural Network Analysis.

Background: Race is commonly used as a proxy for multiple features including socioeconomic status. It is critical to dissociate these factors, to identify mechanisms that affect infant outcomes, such as birth weight, gestational age, and brain development, and to direct appropriate interventions and shape public policy. Methods: Demographic, socioeconomic, and clinical variables were used to model infant outcomes. There were 351 participants included in the analysis for birth weight and gestational age. For the analysis using brain volumes, 280 participants were included after removing participants with missing magnetic resonance imaging scans and those matching our exclusion criteria. We modeled these three different infant outcomes, including infant brain, birth weight, and gestational age, with both linear and nonlinear models. Results: Nonlinear models were better predictors of infant birth weight than linear models (R2 = 0.172 vs. R2 = 0.145, p = .005). In contrast to linear models, nonlinear models ranked income, neighborhood disadvantage, and experiences of discrimination higher in importance than race while modeling birth weight. Race was not an important predictor for either gestational age or structural brain volumes. Conclusions: Consistent with the extant social science literature, the findings related to birth weight suggest that race is a linear proxy for nonlinear factors related to structural racism. Methods that can disentangle factors often correlated with race are important for policy in that they may better identify and rank the modifiable factors that influence outcomes.

Deep Learning Coordinate-Free Quantum Chemistry.

Computing quantum chemical properties of small molecules and polymers can provide insights valuable into physicists, chemists, and biologists when designing new materials, catalysts, biological probes, and drugs. Deep learning can compute quantum chemical properties accurately in a fraction of time required by commonly used methods such as density functional theory. Most current approaches to deep learning in quantum chemistry begin with geometric information from experimentally derived molecular structures or pre-calculated atom coordinates. These approaches have many useful applications, but they can be costly in time and computational resources. In this study, we demonstrate that accurate quantum chemical computations can be performed without geometric information by operating in the coordinate-free domain using deep learning on graph encodings. Coordinate-free methods rely only on molecular graphs, the connectivity of atoms and bonds, without atom coordinates or bond distances. We also find that the choice of graph-encoding architecture substantially affects the performance of these methods. The structures of these graph-encoding architectures provide an opportunity to probe an important, outstanding question in quantum mechanics: what types of quantum chemical properties can be represented by local variable models? We find that Wave, a local variable model, accurately calculates the quantum chemical properties, while graph convolutional architectures require global variables. Furthermore, local variable Wave models outperform global variable graph convolution models on complex molecules with large, correlated systems.

Site-Level Bioactivity of Small-Molecules from Deep-Learned Representations of Quantum Chemistry.

Atom- or bond-level chemical properties of interest in medicinal chemistry, such as drug metabolism and electrophilic reactivity, are important to understand and predict across arbitrary new molecules. Deep learning can be used to map molecular structures to their chemical properties, but the data sets for these tasks are relatively small, which can limit accuracy and generalizability. To overcome this limitation, it would be preferable to model these properties on the basis of the underlying quantum chemical characteristics of small molecules. However, it is difficult to learn higher level chemical properties from lower level quantum calculations. To overcome this challenge, we pretrained deep learning models to compute quantum chemical properties and then reused the intermediate representations constructed by the pretrained network. Transfer learning, in this way, substantially outperformed models based on chemical graphs alone or quantum chemical properties alone. This result was robust, observable in five prediction tasks: identifying sites of epoxidation by metabolic enzymes and identifying sites of covalent reactivity with cyanide, glutathione, DNA and protein. We see that this approach may substantially improve the accuracy of deep learning models for specific chemical structures, such as aromatic systems.