skip to main content

Search for: All records

Award ID contains: 1841520

Note: When clicking on a Digital Object Identifier (DOI) number, you will be taken to an external site maintained by the publisher. Some full text articles may not yet be available without a charge during the embargo (administrative interval).
What is a DOI Number?

Some links on this page may take you to non-federal websites. Their policies may differ from this site.

  1. Xu, Jinbo (Ed.)
    Abstract Motivation Expanding our knowledge of small molecules beyond what is known in nature or designed in wet laboratories promises to significantly advance cheminformatics, drug discovery, biotechnology and material science. In silico molecular design remains challenging, primarily due to the complexity of the chemical space and the non-trivial relationship between chemical structures and biological properties. Deep generative models that learn directly from data are intriguing, but they have yet to demonstrate interpretability in the learned representation, so we can learn more about the relationship between the chemical and biological space. In this article, we advance research on disentangled representation learningmore »for small molecule generation. We build on recent work by us and others on deep graph generative frameworks, which capture atomic interactions via a graph-based representation of a small molecule. The methodological novelty is how we leverage the concept of disentanglement in the graph variational autoencoder framework both to generate biologically relevant small molecules and to enhance model interpretability. Results Extensive qualitative and quantitative experimental evaluation in comparison with state-of-the-art models demonstrate the superiority of our disentanglement framework. We believe this work is an important step to address key challenges in small molecule generation with deep generative frameworks. Availability and implementation Training and generated data are made available at https://ieee-dataport.org/documents/dataset-disentangled-representation-learning-interpretable-molecule-generation. All code is made available at https://anonymous.4open.science/r/D-MolVAE-2799/. Supplementary information Supplementary data are available at Bioinformatics online.« less
    Free, publicly-accessible full text available May 2, 2023
  2. Free, publicly-accessible full text available May 1, 2023
  3. Free, publicly-accessible full text available April 25, 2023
  4. Free, publicly-accessible full text available January 1, 2023
  5. Free, publicly-accessible full text available November 1, 2022
  6. Free, publicly-accessible full text available November 1, 2022
  7. The ocean and atmosphere exert stresses on sea ice that create elongated cracks and leads which dominate the vertical exchange of energy, especially in cold seasons, despite covering only a small fraction of the surface. Motivated by the need of a spatiotemporal analysis of sea ice lead distribution, a practical workflow was developed to classify the high spatial resolution aerial images DMS (Digital Mapping System) along the Laxon Line in the NASA IceBridge Mission. Four sea ice types (thick ice, thin ice, open water, and shadow) were identified, and relevant sea ice lead parameters were derived for the period ofmore »2012–2018. The spatiotemporal variations of lead fraction along the Laxon Line were verified by ATM (Airborne Topographic Mapper) surface height data and correlated with coarse spatial resolution sea ice motion, air temperature, and wind data through multiple regression models. We found that the freeboard data derived from sea ice leads were compatible with other products. The temperature and ice motion vorticity were the leading factors of the formation of sea ice leads, followed by wind vorticity and kinetic moments of ice motion.« less
    Free, publicly-accessible full text available October 1, 2022
  8. Free, publicly-accessible full text available September 1, 2022
  9. COVID-19 has seriously threatened people’s health and well-being across the globe since it was first reported in Wuhan, China in late 2019. This study investigates the mechanism of COVID-19 transmission in different periods within and between cities in China to better understand the nature of the outbreak. We use Moran’s I, a measure of spatial autocorrelation, to examine the spatial dependency of COVID-19 and a dynamic spatial autoregressive model to explore the transmission mechanism. We find that the spatial dependency of COVID-19 decreased over time and that the transmission of the disease could be divided into three distinct stages: anmore »eruption stage, a stabilization stage, and a declination stage. The infection rate between cities was close to one-third of the infection rate within cities at the eruption stage, while it reduced to zero at the declination stage. We also find that the infection rates within cities at the eruption stage and declination stage were similar. China’s policies for controlling the spread of the epidemic, specifically with respect to limiting inter-city mobility and implementing intra-city travel restrictions (social isolation), were most effective in reducing the viral transmission of COVID-19. The findings from this study indicate that the elimination of inter-city mobility had the largest impact on controlling disease transmission.« less