Neural Models for Generating Natural Language Summaries from Temporal Personal Health Data

With an increased interest in the production of personal health technologies designed to track user data (e.g., nutrient  intake, step counts), there is now more opportunity than ever to surface meaningful behavioral insights to everyday users in the form of natural language. This knowledge can increase their behavioral awareness and allow them to take action to meet their health goals. It can also bridge the gap between the vast collection of personal health data and the summary generation required to describe an individual's behavioral tendencies. Previous work has focused on rule-based time-series data summarization methods designed to generate natural language summaries of interesting patterns found within temporal personal health data. We examine recurrent, convolutional, and Transformer-based encoder-decoder models to automatically generate natural language summaries from numeric temporal personal health data. We showcase the effectiveness of our models on real user health data logged in MyFitnessPal (Weber and Achananuparp 2016) and show that we can automatically generate high-quality natural language summaries. Our work serves as a first step towards the ambitious goal of automatically generating novel and meaningful temporal summaries from personal health data.

Demographic and socioeconomic determinants of access to care: A subgroup disparity analysis using new equity-focused measurements

Disparities in healthcare access and utilization associated with demographic and socioeconomic status hinder advancement of health equity. Thus, we designed a novel equity-focused approach to quantify variations of healthcare access/utilization from the expectation in national target populations. We additionally applied survey-weighted logistic regression models, to identify factors associated with usage of a particular type of health care. To facilitate generation of analysis datasets, we built an National Health and Nutrition Examination Survey (NHANES) knowledge graph to help automate source-level dynamic analyses across different survey years and subjects' characteristics. We performed a cross-sectional subgroup disparity analysis of 2013-2018 NHANES on U.S. adults for receipt of diabetes treatments and vaccines against Hepatitis A (HAV), Hepatitis B (HBV), and Human Papilloma (HPV). Results show that in populations with hemoglobin A1c level ≥6%, patients with non-private insurance were less likely to receive newer and more beneficial antidiabetic medications; being Asian further exacerbated these disparities. For widely used drugs such as insulin, Asians experienced insignificant disparities in odds of prescription compared to White patients but received highly inadequate treatments with regard to their distribution in U.S. diabetic population. Vaccination rates were associated with some demographic/socioeconomic factors but not the others at different degrees for different diseases. For instance, while equity scores increase with rising education levels for HBV, they decrease with rising wealth levels for HPV. Among women vaccinated against HPV, minorities and poor communities usually received Cervarix while non-Hispanic White and higher-income groups received the more comprehensive Gardasil vaccine. Our study identified and quantified the impact of determinants of healthcare utilization for antidiabetic medications and vaccinations. Our new methods for semantics-aware disparity analysis of NHANES data could be readily generalized to other public health goals to support more rapid identification of disparities and development of policies, thus advancing health equity.

Structure-informed clustering for population stratification in association studies

BACKGROUND

Identifying variants associated with complex traits is a challenging task in genetic association studies due to linkage disequilibrium (LD) between genetic variants and population stratification, unrelated to the disease risk. Existing methods of population structure correction use principal component analysis or linear mixed models with a random effect when modeling associations between a trait of interest and genetic markers. However, due to stringent significance thresholds and latent interactions between the markers, these methods often fail to detect genuinely associated variants.

RESULTS

To overcome this, we propose CluStrat, which corrects for complex arbitrarily structured populations while leveraging the linkage disequilibrium induced distances between genetic markers. It performs an agglomerative hierarchical clustering using the Mahalanobis distance covariance matrix of the markers. In simulation studies, we show that our method outperforms existing methods in detecting true causal variants. Applying CluStrat on WTCCC2 and UK Biobank cohorts, we found biologically relevant associations in Schizophrenia and Myocardial Infarction. CluStrat was also able to correct for population structure in polygenic adaptation of height in Europeans.

CONCLUSIONS

CluStrat highlights the advantages of biologically relevant distance metrics, such as the Mahalanobis distance, which captures the cryptic interactions within populations in the presence of LD better than the Euclidean distance.

Open Macromolecular Genome: Generative Design of Synthetically Accessible Polymers

A grand challenge in polymer science lies in the predictive design of new polymeric materials with targeted functionality. However, de novo design of functional polymers is challenging due to the vast chemical space and an incomplete understanding of structure-property relations. Recent advances in deep generative modeling have facilitated the efficient exploration of molecular design space, but data sparsity in polymer science is a major obstacle hindering progress. In this work, we introduce a vast polymer database known as the Open Macromolecular Genome (OMG), which contains synthesizable polymer chemistries compatible with known polymerization reactions and commercially available reactants selected for synthetic feasibility. The OMG is used in concert with a synthetically aware generative model known as Molecule Chef to identify property-optimized constitutional repeating units, constituent reactants, and reaction pathways of polymers, thereby advancing polymer design into the realm of synthetic relevance. As a proof-of-principle demonstration, we show that polymers with targeted octanol-water solubilities are readily generated together with monomer reactant building blocks and associated polymerization reactions. Suggested reactants are further integrated with Reaxys polymerization data to provide hypothetical reaction conditions (e.g., temperature, catalysts, and solvents). Broadly, the OMG is a polymer design approach capable of enabling data-intensive generative models for synthetic polymer design. Overall, this work represents a significant advance, enabling the property targeted design of synthetic polymers subject to practical synthetic constraints.

Trade-off between Gradual Set and On/Off Ratio in HfOx-Based Analog Memory with a Thin SiOx Barrier Layer

HfO_x-based synapses are widely accepted as a viable candidate for both in-memory and neuromorphic computing. Resistance change in oxide-based synapses is caused by the motion of oxygen vacancies. HfO_x-based synapses typically demonstrate an abrupt nonlinear resistance change under positive bias application (set), limiting their viability as analog memory. In this work, a thin barrier layer of AlO_x or SiO_x is added to the bottom electrode/oxide interface to slow the migration of oxygen vacancies. Electrical results show that the resistance change in HfO_x/SiO_x devices is more controlled than the HfO_x devices during the set. While the on/off ratio for the HfO_x/SiO_x devices is still large (∼10), it is shown to be smaller than that of HfO_x/AlO_x and HfO_x devices. Finite element modeling suggests that the slower oxygen vacancy migration in HfO_x/SiO_x devices during reset results in a narrower rupture region in the conductive filament. The narrower rupture region causes a lower high resistance state and, thus, a smaller on/off ratio for the HfO_x/SiO_x devices. Overall, the results show that slowing the motion of oxygen vacancies in the barrier layer devices improves the resistance change during the set but lowers the on/off ratio.

B-cell epitope discovery: The first protein flexibility-based algorithm-Zika virus conserved epitope demonstration

Antibody-antigen interaction-at antigenic local environments called B-cell epitopes-is a prominent mechanism for neutralization of infection. Effective mimicry, and display, of B-cell epitopes is key to vaccine design. Here, a physical approach is evaluated for the discovery of epitopes which evolve slowly over closely related pathogens (conserved epitopes). The approach is 1) protein flexibility-based and 2) demonstrated with clinically relevant enveloped viruses, simulated via molecular dynamics. The approach is validated against 1) seven structurally characterized enveloped virus epitopes which evolved the least (out of thirty-nine enveloped virus-antibody structures), 2) two structurally characterized non-enveloped virus epitopes which evolved slowly (out of eight non-enveloped virus-antibody structures), and 3) eight preexisting epitope and peptide discovery algorithms. Rationale for a new benchmarking scheme is presented. A data-driven epitope clustering algorithm is introduced. The prediction of five Zika virus epitopes (for future exploration on recombinant vaccine technologies) is demonstrated. For the first time, protein flexibility is shown to outperform solvent accessible surface area as an epitope discovery metric.

Discovery of C20-Diterpenoid Alkaloid Kobusine Derivatives Exhibiting Sub-G1 Inducing Activity

Although many diterpenoid alkaloids have been evaluated recently for antiproliferative activity against human cancer cell lines, little data have been offered relating to the antiproliferative effects of hetisine-type C₂₀-diterpenoid alkaloids, such as kobusine (1), likewise as their derivatives. A total of 43 novel diterpenoid alkaloid derivatives (2-10, 2b, 3a, 3b, 6a-16a, 7b, 9b, 10b, 13, 15-26, 15b, 18a, 23a, 27a) were prepared by C-11 and -15 esterification of 1. Antiproliferative effects of the natural parent compound (1) and all synthesized kobusine derivatives against human cancer cell lines, including a triple-negative breast cancer (TNBC) cell line as well as a P-glycoprotein overexpressing multidrug-resistant subline, were assessed. The structure-based design strategy resulted in the lead derivative 11,15-dibenzoylkobusine (3; average IC₅₀ 7.3 μM). Several newly synthesized kobusine derivatives (particularly, 5-8, 10, 13, 15-26) exhibited substantial suppressive effects against all tested human cancer cell lines. In contrast, kobusine (1), 11,15-O-diacetylkobusine (2), 11-acylkobusine derivatives (3a, 6a, 9a, 11a, 12a, 15a, 27a), and 15-acylkobusine derivatives (2b, 3b, 7b, 9b, 10b, 15b) showed no effect. The most active kobusine derivatives primarily had two specific substitution patterns, C-11,15 and C-11. Notably, 11,15-diacylkobusine derivatives (3, 6-10, 13, 15, 16, 18, 23) were more potent compared with 11- and 15-acylkobusine derivatives (3a, 3b, 6a-10a, 7b, 9b, 10b, 13a, 15a, 15b, 16a, 18a, 23a). Derivatives 13 and 25 induced MDA-MB-231 cells to the sub-G1 phase within 12 h. 11,15-Diacylation of kobusine (1) appears to be crucial for inducing antiproliferative activity in this alkaloid class and could introduce a new avenue to overcome TNBC using natural product derivatives.

Domain Generalization for Language-Independent Automatic Speech Recognition

A language-independent automatic speech recognizer (ASR) is one that can be used for phonetic transcription in languages other than the languages in which it was trained. Language-independent ASR is difficult to train, because different languages implement phones differently: even when phonemes in two different languages are written using the same symbols in the international phonetic alphabet, they are differentiated by different distributions of language-dependent redundant articulatory features. This article demonstrates that the goal of language-independence may be approximated in different ways, depending on the size of the training set, the presence vs. absence of familial relationships between the training and test languages, and the method used to implement phone recognition or classification. When the training set contains many languages, and when every language in the test set is related (shares the same language family with) a language in the training set, then language-independent ASR may be trained using an empirical risk minimization strategy (e.g., using connectionist temporal classification without extra regularizers). When the training set is limited to a small number of languages from one language family, however, and the test languages are not from the same language family, then the best performance is achieved by using domain-invariant representation learning strategies. Two different representation learning strategies are tested in this article: invariant risk minimization, and regret minimization. We find that invariant risk minimization is better at the task of phone token classification (given known segment boundary times), while regret minimization is better at the task of phone token recognition.

Accelerating high-throughput virtual screening through molecular pool-based active learning

Structure-based virtual screening is an important tool in early stage drug discovery that scores the interactions between a target protein and candidate ligands. As virtual libraries continue to grow (in excess of 108 molecules), so too do the resources necessary to conduct exhaustive virtual screening campaigns on these libraries. However, Bayesian optimization techniques, previously employed in other scientific discovery problems, can aid in their exploration: a surrogate structure-property relationship model trained on the predicted affinities of a subset of the library can be applied to the remaining library members, allowing the least promising compounds to be excluded from evaluation. In this study, we explore the application of these techniques to computational docking datasets and assess the impact of surrogate model architecture, acquisition function, and acquisition batch size on optimization performance. We observe significant reductions in computational costs; for example, using a directed-message passing neural network we can identify 94.8% or 89.3% of the top-50 000 ligands in a 100M member library after testing only 2.4% of candidate ligands using an upper confidence bound or greedy acquisition strategy, respectively. Such model-guided searches mitigate the increasing computational costs of screening increasingly large virtual libraries and can accelerate high-throughput virtual screening campaigns with applications beyond docking.

The IMEx coronavirus interactome: an evolving map of Coronaviridae-host molecular interactions

The current coronavirus disease of 2019 (COVID-19) pandemic, caused by the severe acute respiratory syndrome coronavirus (SARS-CoV)-2, has spurred a wave of research of nearly unprecedented scale. Among the different strategies that are being used to understand the disease and develop effective treatments, the study of physical molecular interactions can provide fine-grained resolution of the mechanisms behind the virus biology and the human organism response. We present a curated dataset of physical molecular interactions focused on proteins from SARS-CoV-2, SARS-CoV-1 and other members of the Coronaviridae family that has been manually extracted by International Molecular Exchange (IMEx) Consortium curators. Currently, the dataset comprises over 4400 binarized interactions extracted from 151 publications. The dataset can be accessed in the standard formats recommended by the Proteomics Standards Initiative (HUPO-PSI) at the IntAct database website (https://www.ebi.ac.uk/intact) and will be continuously updated as research on COVID-19 progresses.

Challenges in the construction of knowledge bases for human microbiome-disease associations

The last few years have seen tremendous growth in human microbiome research, with a particular focus on the links to both mental and physical health and disease. Medical and experimental settings provide initial sources of information about these links, but individual studies produce disconnected pieces of knowledge bounded in context by the perspective of expert researchers reading full-text publications. Building a knowledge base (KB) consolidating these disconnected pieces is an essential first step to democratize and accelerate the process of accessing the collective discoveries of human disease connections to the human microbiome. In this article, we survey the existing tools and development efforts that have been produced to capture portions of the information needed to construct a KB of all known human microbiome-disease associations and highlight the need for additional innovations in natural language processing (NLP), text mining, taxonomic representations, and field-wide vocabulary standardization in human microbiome research. Addressing these challenges will enable the construction of KBs that help identify new insights amenable to experimental validation and potentially clinical decision support.

Recruiting patients as partners in health research: a qualitative descriptive study

PLAIN ENGLISH SUMMARY

Increasingly, funders and researchers want to partner with patients in health research, but it can be challenging for researchers to find patient partners. More than taking part in research as participants, patient partners help design, carry out and manage research projects. The goal of this study was to describe ways that patient partners have been recruited by researchers and patient engagement leads (individuals within organizations responsible for promoting and supporting patients as research partners). We talked with researchers and patient engagement leads in Canada and the United Kingdom, as well as a patient representative. We found three ways that could help researchers and patients find each other. One way is a case-by-case basis, where patients are often sought with experience of a health condition that is the focus of the research. The other ways involved directories where projects were posted and could be found by patients and researchers, or a third party matched patients with research projects. We found four recruitment strategies:Social marketingCommunity outreachHealth systemPartnering with other organizations (e.g., advocacy groups) There are many influences on finding, selecting and retaining patient partners: patient characteristics, the local setting, the opportunity, work climate, education and support. We hope study results will provide a useful starting point for research teams in recruiting their patient partners.

ABSTRACT

Background Patient engagement in clinical trials and other health research continues to gain momentum. While the benefits of patient engagement in research are emerging, relatively little is known about recruiting patients as research partners. The purpose of this study was to describe recruitment strategies and models of recruiting patients as partners in health research. Methods Qualitative descriptive study. Thirteen patient engagement leads and health researchers from Canada and the United Kingdom, as well as one patient representative from a national patient organization (7 female) completed semi-structured interviews. Results Recruitment infrastructures available to respondents varied, but could be categorized into three models including the traditional, third-party and directory models. Four categories of recruitment strategies were identified, representing multiple ways of recruiting patient partners: social marketing recruitment, community outreach recruitment, health system recruitment, and partnering recruitment. Conclusions Multiple recruitment strategies were identified for engaging patient partners in research, and some common factors influenced recruitment. Study findings contribute to the evidence base in patient engagement and provide guidance for research teams to help identify potential recruitment methods for their patient partners.

Butterfly magnetoresistance, quasi-2D Dirac Fermi surface and topological phase transition in ZrSiS

Magnetoresistance (MR), the change of a material's electrical resistance in response to an applied magnetic field, is a technologically important property that has been the topic of intense study for more than a quarter century. We report the observation of an unusual "butterfly"-shaped titanic angular magnetoresistance (AMR) in the nonmagnetic Dirac material, ZrSiS, which we find to be the most conducting sulfide known, with a 2-K resistivity as low as 48(4) nΩ⋅cm. The MR in ZrSiS is large and positive, reaching nearly 1.8 × 10⁵ percent at 9 T and 2 K at a 45° angle between the applied current (I || a) and the applied field (90° is H || c). Approaching 90°, a "dip" is seen in the AMR, which, by analyzing Shubnikov de Haas oscillations at different angles, we find to coincide with a very sharp topological phase transition unlike any seen in other known Dirac/Weyl materials. We find that ZrSiS has a combination of two-dimensional (2D) and 3D Dirac pockets comprising its Fermi surface and that the combination of high-mobility carriers and multiple pockets in ZrSiS allows for large property changes to occur as a function of angle between applied fields. This makes it a promising platform to study the physics stemming from the coexistence of 2D and 3D Dirac electrons as well as opens the door to creating devices focused on switching between different parts of the Fermi surface and different topological states.