The Automatic Context Measurement Tool (ACMT) to Compile Participant-Specific Built and Social Environment Measures for Health Research: Development and Usability Study.

BACKGROUND: The environment shapes health behaviors and outcomes. Studies exploring this influence have been limited to research groups with the geographic information systems expertise required to develop built and social environment measures (eg, groups that include a researcher with geographic information system expertise). OBJECTIVE: The goal of this study was to develop an open-source, user-friendly, and privacy-preserving tool for conveniently linking built, social, and natural environmental variables to study participant addresses. METHODS: We built the automatic context measurement tool (ACMT). The ACMT comprises two components: (1) a geocoder, which identifies a latitude and longitude given an address (currently limited to the United States), and (2) a context measure assembler, which computes measures from publicly available data sources linked to a latitude and longitude. ACMT users access both of these components using an RStudio/RShiny-based web interface that is hosted within a Docker container, which runs on a local computer and keeps user data stored in local to protect sensitive data. We illustrate ACMT with 2 use cases: one comparing population density patterns within several major US cities, and one identifying correlates of cannabis licensure status in Washington State. RESULTS: In the population density analysis, we created a line plot showing the population density (x-axis) in relation to distance from the center of the city (y-axis, using city hall location as a proxy) for Seattle, Los Angeles, Chicago, New York City, Nashville, Houston, and Boston with the distances being 1000, 2000, 3000, 4000, and 5000 m. We found the population density tended to decrease as distance from city hall increased except for Nashville and Houston, 2 cities that are notably more sprawling than the others. New York City had a significantly higher population density than the others. We also observed that Los Angeles and Seattle had similarly low population densities within up to 2500 m of City Hall. In the cannabis licensure status analysis, we gathered neighborhood measures such as age, sex, commute time, and education. We found the strongest predictive characteristic of cannabis license approval to be the count of female children aged 5 to 9 years and the proportion of females aged 62 to 64 years who were not in the labor force. However, after accounting for Bonferroni error correction, none of the measures were significantly associated with cannabis retail license approval status. CONCLUSIONS: The ACMT can be used to compile environmental measures to study the influence of environmental context on population health. The portable and flexible nature of ACMT makes it optimal for neighborhood study research seeking to attribute environmental data to specific locations within the United States.

Milk proteomic analysis reveals differentially expressed proteins in high-yielding and low-yielding Guanzhong dairy goats at peak lactation.

The aim of this experiment was to investigate the differential proteomic characteristics of milk from high- and low-yielding Guanzhong dairy goats during the peak lactation period under the same feeding conditions. Nine Guanzhong dairy goats with high yield (H: 3.5 ± 0.17 kg/d) and nine with low yield (L:1.2 ± 0.25 kg/d) were selected for milk proteomic analysis using tandem mass tag technology. A total of 78 differentially expressed proteins were identified. Compared with L, 50 proteins including HK3, HSPB1 and ANXA2 were significantly upregulated in H milk, while 28 proteins including LALBA and XDH were significantly downregulated. Bioinformatics analysis of the differentially expressed proteins showed that galactose metabolism, purine metabolism, glycolysis/gluconeogenesis, MAPK signaling pathway, regulation of actin cytoskeleton and other pathways were closely related to milk yield. HK3, HSPB1, ANXA2, LALBA and XDH were important candidate proteins associated with the milk production characteristics of Guanzhong dairy goats. Our data provide relevant biomarkers and a theoretical basis for improving milk production in Guanzhong dairy goats.

Comparative analysis of different metabolites in semen of Guanzhong dairy goats with different motility rates.

In order to explore the different metabolites of buck semen with different motility stored at 4 °C, the semen of bucks was collected by artificial vagina. The collected semen was divided into high motility group and low motility group after treatment, with 6 replicates set for each group. The semen metabolites of high motility group and low motility group were detected by Liquid Chromatography-Mass Spectrometry (LC-MS). The results showed that 101 different metabolites were detected in the high and low motility groups of bucks, of which 48 metabolites were significantly up-regulated (P < 0.05) and 53 metabolites were significantly down regulated (P < 0.05). Most of these metabolites belonged to lipids and lipid-like molecules, organic acids and their derivatives, and organic oxygen compounds, which were mainly related to energy metabolism. According to the functional enrichment analysis of the former differential metabolites in KEGG database, the top 20 most representative metabolic pathways were detected, among which the glycerophospholipid metabolic pathways changed significantly. From the perspective of metabolomics, this study revealed the differences of metabolites and characteristic compounds of semen with different motility of bucks under low temperature preservation, which provided a scientific basis for the preservation and utilization of semen of Guanzhong dairy goats in the future.

Data Augmentation in Defect Detection of Sanitary Ceramics in Small and Non-i.i.d Datasets.

In this study, a data-augmentation method is proposed to narrow the significant difference between the distribution of training and test sets when small sample sizes are concerned. Two major obstacles exist in the process of defect detection on sanitary ceramics. The first results from the high cost of sample collection, namely, the difficulty in obtaining a large number of training images required by deep-learning algorithms, which limits the application of existing algorithms in sanitary-ceramic defect detection. Second, due to the limitation of production processes, the collected defect images are often marked, thereby resulting in great differences in distribution compared with the images of test sets, which further affects the performance of detect-detection algorithms. The lack of training data and the differences in distribution between training and test sets lead to the fact that existing deep learning-based algorithms cannot be used directly in the defect detection of sanitary ceramics. The method proposed in this study, which is based on a generative adversarial network and the Gaussian mixture model, can effectively increase the number of training samples and reduce distribution differences between training and test sets, and the features of the generated images can be controlled to a certain extent. By applying this method, the accuracy is improved from approximately 75% to nearly 90% in almost all experiments on different classification networks.

Deep Representations of First-person Pronouns for Prediction of Depression Symptom Severity.

Prior work has shown that analyzing the use of first-person singular pronouns can provide insight into individuals' mental status, especially depression symptom severity. These findings were generated by counting frequencies of first-person singular pronouns in text data. However, counting doesn't capture how these pronouns are used. Recent advances in neural language modeling have leveraged methods generating contextual embeddings. In this study, we sought to utilize the embeddings of first-person pronouns obtained from contextualized language representation models to capture ways these pronouns are used, to analyze mental status. De-identified text messages sent during online psychotherapy with weekly assessment of depression severity were used for evaluation. Results indicate the advantage of contextualized first-person pronoun embeddings over standard classification token embeddings and frequency-based pronoun analysis results in predicting depression symptom severity. This suggests contextual representations of first-person pronouns can enhance the predictive utility of language used by people with depression symptoms.