California Exodus? A Network Model of Population Redistribution in the United States.

Motivated by debates about California's net migration loss, we employ valued exponential-family random graph models to analyze the inter-county migration flow networks in the United States. We introduce a protocol that visualizes the complex effects of potential underlying mechanisms, and perform in silico knockout experiments to quantify their contribution to the California Exodus. We find that racial dynamics contribute to the California Exodus, urbanization ameliorates it, and political climate and housing costs have little impact. Moreover, the severity of the California Exodus depends on how one measures it, and California is not the state with the most substantial population loss. The paper demonstrates how generative statistical models can provide mechanistic insights beyond simple hypothesis-testing.

Malakoplakia Involving the Maxilla: A Case Report and a Review of the Literature.

Malakoplakia is a rare inflammatory disorder which typically occurs in immunocompromised patients secondary to impaired bactericidal activity of macrophages. While this entity commonly arises in the genitourinary and gastrointestinal tracts, lesions of the head and neck have been reported only rarely, with oral cavity involvement reported in 3 cases. The most common presentation of head and neck malakoplakia is that of a cutaneous flesh-colored papule or nodule. This case report, however, illustrates the first time malakoplakia is identified affecting the maxilla and maxillary alveolar ridge mucosa. Histochemical and immunohistochemical stains are presented and include positivity for PAS, von Kossa stain, iron stain, and CD68 and negativity for GMS and Gram stains, indicating an inability to demonstrate microbial infection. Thus, clinicians and pathologists alike should be aware of malakoplakia as a pathologic entity when forming differential diagnoses, particularly in immunosuppressed individuals.

Continuous Time Graph Processes with Known ERGM Equilibria: Contextual Review, Extensions, and Synthesis.

Graph processes that unfold in continuous time are of obvious theoretical and practical interest. Particularly useful are those whose long-term behavior converges to a graph distribution of known form. Here, we review some of the conditions for such convergence, and provide examples of novel and/or known processes that do so. These include subfamilies of the well-known stochastic actor oriented models, as well as continuum extensions of temporal and separable temporal exponential family random graph models. We also comment on some related threads in the broader work on network dynamics, which provide additional context for the continuous time case. Graph processes that unfold in continuous time are natural models for social network dynamics: able to directly represent changes in structure as they unfold (rather than, e.g. as snapshots at discrete intervals), such models not only offer the promise of capturing dynamics at high temporal resolution, but are also easily mapped to empirical data without the need to preselect a level of granularity with respect to which the dynamics are defined. Although relatively few general frameworks of this type have been extensively studied, at least one (the stochastic actor-oriented models, or SAOMs) is arguably among the most successful and widely used families of models in the social sciences (see, e.g., Snijders (2001); Steglich et al. (2010); Burk et al. (2007); Sijtsema et al. (2010); de la Haye et al. (2011); Weerman (2011); Schaefer and Kreager (2020) among many others). Work using other continuous time graph processes has also found applications both within (Koskinen and Snijders, 2007; Koskinen et al., 2015; Stadtfeld et al., 2017; Hoffman et al., 2020) and beyond (Grazioli et al., 2019; Yu et al., 2020) the social sciences, suggesting the potential for further advances.

Challenges for Nurse Anesthetists Reentering Practice Following Substance Use Disorder Treatment.

Substance use disorder (SUD) is a persistent, relapsing condition that is present in approximately 10% of anesthesia providers, who, compared with other healthcare providers, face a greater risk of developing an SUD by virtue of constant access to medications. The ability of certified registered nurse anesthesiologists (CRNAs) to obtain or maintain employment after treatment for SUD treatment is not well documented. The purpose of this qualitative study was to explore challenges encountered by CRNAs in recovery as they attempt to reenter practice following SUD treatment. The phenomenon was explored through multiple-case study, using qualitative semistructured interviews with participants in four cases: CRNAs in recovery, CRNA colleagues, CRNA employers, and professional health program employees. Thirty-six participants conveyed their perspectives about challenges that CRNAs in recovery face upon reentry into practice following SUD treatment. The Worker Well-Being conceptual model was used to guide this study. The study revealed that more SUD education is a key facilitator for reentry, risk of relapse was a major concern, and stigma was the most significant barrier for CRNAs in recovery. Stigma persists as a considerable barrier in many facets of SUD, contributing to an increase in shame associated with having the disease.

Dysgenetic Polycystic Disease of the Salivary Glands: A Case Report of This Rare Entity Occurring for the First Time in the Minor Salivary Glands of the Tongue, and a Review of the Literature.

Dysgenetic polycystic disease, also known just as polycystic disease, is a very rare developmental abnormality affecting the salivary gland duct system. This entity has been reported in only 21 patients previously, although a careful review suggests only 16 patients have histological evidence of the disease. In previously reported cases, this lesion most commonly presents as either an incidental finding or as a swelling affecting the parotid glands bilaterally, or rarely the submandibular glands bilaterally. This case report details the first time dysgenetic polycystic disease is found affecting the minor salivary glands of the tongue in a 55-year-old male. Histochemical and immunohistochemical stains are presented and include positivity for AE1/AE3 and p63, and negativity for progesterone receptor, androgen receptor, mammaglobin, S100 and BRAF V600E. PAS-D and Congo Red highlight special microamyloid spheroliths structures intraluminally.

Environmental and geographical factors influence the occurrence and abundance of the southern house mosquito, Culex quinquefasciatus, in Hawai'i.

Hawaiian honeycreepers, a group of endemic Hawaiian forest birds, are being threatened by avian malaria, a non-native disease that is driving honeycreepers populations to extinction. Avian malaria is caused by the parasite Plasmodium relictum, which is transmitted by the invasive mosquito Culex quinquefasciatus. Environmental and geographical factors play an important role in shaping mosquito-borne disease transmission dynamics through their influence on the distribution and abundance of mosquitoes. We assessed the effects of environmental (temperature, precipitation), geographic (site, elevation, distance to anthropogenic features), and trap type (CDC light trap, CDC gravid trap) factors on mosquito occurrence and abundance. Occurrence was analyzed using classification and regression tree models (CART) and generalized linear models (GLM); abundance (count data) was analyzed using generalized linear mixed models (GLMMs). Models predicted highest mosquito occurrence at mid-elevation sites and between July and November. Occurrence increased with temperature and precipitation up to 580 mm. For abundance, the best model was a zero-inflated negative-binomial model that indicated higher abundance of mosquitoes at mid-elevation sites and peak abundance between August and October. Estimation of occurrence and abundance as well as understanding the factors that influence them are key for mosquito control, which may reduce the risk of forest bird extinction.

The HER2-directed antibody-drug conjugate DHES0815A in advanced and/or metastatic breast cancer: preclinical characterization and phase 1 trial results.

Approved antibody-drug conjugates (ADCs) for HER2-positive breast cancer include trastuzumab emtansine and trastuzumab deruxtecan. To develop a differentiated HER2 ADC, we chose an antibody that does not compete with trastuzumab or pertuzumab for binding, conjugated to a reduced potency PBD (pyrrolobenzodiazepine) dimer payload. PBDs are potent cytotoxic agents that alkylate and cross-link DNA. In our study, the PBD dimer is modified to alkylate, but not cross-link DNA. This HER2 ADC, DHES0815A, demonstrates in vivo efficacy in models of HER2-positive and HER2-low cancers and is well-tolerated in cynomolgus monkey safety studies. Mechanisms of action include induction of DNA damage and apoptosis, activity in non-dividing cells, and bystander activity. A dose-escalation study (ClinicalTrials.gov: NCT03451162) in patients with HER2-positive metastatic breast cancer, with the primary objective of evaluating the safety and tolerability of DHES0815A and secondary objectives of characterizing the pharmacokinetics, objective response rate, duration of response, and formation of anti-DHES0815A antibodies, is reported herein. Despite early signs of anti-tumor activity, patients at higher doses develop persistent, non-resolvable dermal, ocular, and pulmonary toxicities, which led to early termination of the phase 1 trial.

Suicide rates in the UK Armed Forces, compared with the general workforce and merchant shipping during peacetime years since 1900.

INTRODUCTION: The main objective was to compare suicide rates and their trends across the three UK Armed forces (Royal Navy, Army and Royal Air Force) from 1900 to 2020. Further objectives were to compare suicide rates with those in the corresponding general population and in UK merchant shipping and to discuss preventative measures. METHODS: Examination of annual mortality reports and returns, death inquiry files and official statistics. The main outcome measure was the suicide rate per 100 000 population employed. RESULTS: Since 1990, there have been significant reductions in suicide rates in each of the Armed Forces, although a non-significant increase in the Army since 2010. Compared with the corresponding general population, during the most recent decade from 2010 up to 2020, suicide rates were 73% lower in the Royal Air Force, 56% lower in the Royal Navy and 43% lower in the Army. Suicide rates have been significantly decreased in the Royal Air Force since the 1950s, in the Royal Navy since the 1970s and in the Army since the 1980s (comparisons for the Royal Navy and the Army were not available from the late 1940s to the 1960s).During the earliest decades from 1900 to the 1930s, suicide rates in the Armed Forces were mostly quite similar or moderately increased compared with the general population, but far lower than in merchant shipping. Following legislative changes in the last 30 years, suicide rates through poisoning by gases and through firearms or explosives have fallen sharply. CONCLUSIONS: The study shows that suicide rates in the Armed Forces have been lower than in the general population over many decades. The sharp reductions in suicide rates over the last 30 years suggest the effectiveness of recent preventative measures, including reductions in access to a method of suicide and well-being initiatives.

Linking avian malaria parasitemia estimates from quantitative PCR and microscopy reveals new infection patterns in Hawai'i.

Plasmodium parasites infect thousands of species and provide an exceptional system for studying host-pathogen dynamics, especially for multi-host pathogens. However, understanding these interactions requires an accurate assay of infection. Assessing Plasmodium infections using microscopy on blood smears often misses infections with low parasitemias (the fractions of cells infected), and biases in malaria prevalence estimates will differ among hosts that differ in mean parasitemias. We examined Plasmodium relictum infection and parasitemia using both microscopy of blood smears and quantitative polymerase chain reaction (qPCR) on 299 samples from multiple bird species in Hawai'i and fit models to predict parasitemias from qPCR cycle threshold (Ct) values. We used these models to quantify the extent to which microscopy underestimated infection prevalence and to more accurately estimate infection patterns for each species for a large historical study done by microscopy. We found that most qPCR-positive wild-caught birds in Hawaii had low parasitemias (Ct scores ≥35), which were rarely detected by microscopy. The fraction of infections missed by microscopy differed substantially among eight species due to differences in species' parasitemia levels. Infection prevalence was likely 4-5-fold higher than previous microscopy estimates for three introduced species, including Zosterops japonicus, Hawaii's most abundant forest bird, which had low average parasitemias. In contrast, prevalence was likely only 1.5-2.3-fold higher than previous estimates for Himatione sanguinea and Chlorodrepanis virens, two native species with high average parasitemias. Our results indicate that relative patterns of infection among species differ substantially from those observed in previous microscopy studies, and that differences depend on variation in parasitemias among species. Although microscopy of blood smears is useful for estimating the frequency of different Plasmodium stages and host attributes, more sensitive quantitative methods, including qPCR, are needed to accurately estimate and compare infection prevalence among host species.

CD1a is not effective in distinguishing calcifying epithelial odontogenic tumors from amyloid-rich central odontogenic fibroma.

OBJECTIVE: Differential diagnosis between the non-calcifying variant of calcifying epithelial odontogenic tumor (NCLC-CEOT) and amyloid-rich central odontogenic fibroma (AR-COdF) has become a debate, particularly regarding the frequency of CD1a positivity in both entities. This has led to the growing consensus that CD1a-positive staining in AR-NC lesions confirms the diagnosis of AR-COdF. Here, we assess the validity of this consensus. STUDY DESIGN: We collected the data of a case series of histopathologically distinct CEOTs, NCLC-CEOTs, and COdFs and stained them for CD1a and amyloid. Of the 9 CEOTs and NCLC-CEOTs, we diagnosed 4 as classic, 3 as associated with a dentigerous cyst, and 2 as combined CEOT/adenomatoid odontogenic tumors. Of the 9 COdFs, we diagnosed 3 as epithelial poor, 3 as epithelial rich (lacking amyloid), 2 as hyalinized with amyloid, and 1 as hyalinized without amyloid and assessed the staining results. RESULTS: Of the 9 CEOTs and NCLC-CEOTs, 7 stained positively for CD1a, 5 diffusely and 2 focally. Notably, 2 classic NCLC-CEOTs stained strongly CD1a positive. All 3 of the epithelial-poor COdFs were predictably CD1a negative. Of the 6 remaining COdFs, 2 were CD1a positive, 1 hyalinized-with-amyloid COdF diffusely and 1 epithelial-rich-without amyloid focally. CONCLUSIONS: CD1a positivity, which occurs in classic CEOT and NCLC-CEOT, does not help distinguish between NCLC-CEOT and AR-COdF and is inconsistent in all AR-COdFs. The diagnosis of CEOT and AR-COdF should be guided by appropriate histopathologic criteria irrespective of CD1a staining or the presence of amyloid or calcifications.

Spatially resolved detection of small molecules from press-dried plant tissue using MALDI imaging.

Premise: Matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI-MSI) is a chemical imaging method that can visualize spatial distributions of particular molecules. Plant tissue imaging has so far mostly used cryosectioning, which can be impractical for the preparation of large-area imaging samples, such as full flower petals. Imaging unsectioned plant tissue presents its own difficulties in extracting metabolites to the surface due to the waxy cuticle. Methods: We address this by using established delipidation techniques combined with a solvent vapor extraction prior to applying the matrix with many low-concentration sprays. Results: Using this procedure, we imaged tissue from three different plant species (two flowers and one carnivorous plant leaf). Material factorization analysis of the resulting data reveals a wide range of plant-specific small molecules with varying degrees of localization to specific portions of the tissue samples, while facilitating detection and removal of signal from background sources. Conclusions: This work demonstrates applicability of MALDI-MSI to press-dried plant samples without freezing or cryosectioning, setting the stage for spatially resolved molecule identification. Increased mass resolution and inclusion of tandem mass spectrometry are necessary next steps to allow more specific and reliable compound identification.

Premisa: Matrix­assisted laser desorption/ionization mass spectrometry imaging (MALDI­MSI) es un método de imagen química que puede visualizar distribuciones espaciales de moléculas particulares. Hasta ahora, las imágenes de tejido vegetal han utilizado principalmente la criosección, lo cual puede ser poco práctico para la preparación de muestras de imágenes con áreas grandes, tales como los pétalos completos de una flor. La obtención de imágenes de tejido vegetal no seccionado presenta sus propias dificultades durante la extracción de metabolitos a la superficie, debido a la cutícula cerosa de la planta. Métodos: Abordamos esto usando técnicas establecidas de deslipidación combinados con una extracción de vapor por solvente antes de aplicar por aspersión la matriz en bajas concentraciones. Resultados: Usando este procedimiento, obtuvimos imágenes de tejido de tres especies de plantas diferentes (dos flores y una hoja de planta carnívora). Análisis de factorización material de los datos obtenidos revelaron una amplia gama de pequeñas moléculas específicas en plantas con diversos grados de localización en porciones específicas de las muestras de tejido, al igual que facilitó la detección y remoción de las señales de fondo. Conclusión: Nuestro trabajo demuestra la aplicabilidad de MALDI­MSI hacía muestras de plantas secadas a presión sin congelación o criosección, creando el marco para la identificación de moléculas resueltas espacialmente. Aumento de la resolución de masas e inclusión de la espectrometría de masas en tándem son pasos necesarios para obtener identificación de compuestos más específica y confiable.

The virtual knee clinic - A tool to streamline new outpatient referrals.

INTRODUCTION: Traditionally it has been the case for orthopaedic consultants to review GP referrals for the orthopaedic outpatient clinic where possible in amongst other clinical commitments. This could sometimes lead to unsuitable patients being reviewed and both patients and clinicians becoming frustrated. Building on the virtual fracture clinic, a new screening tool was implemented to streamline new referrals. The aim of this study is to investigate the change in patients given outpatient appointments following the introduction of a new streamlining protocol. METHODS: Referrals had to meet the criteria of BMI under 40 or evidence of weight loss effort, recent radiographs and appropriate clinical details in keeping with Getting It Right First Time (GIRFT). Consultant were given dedicated clinical time to review and either triage the patient to the most appropriate clinic type, or return the referral with advice to the GP. 10 months of data was collected prior to the protocol and 10 months after implementation. RESULTS: 1781 patients were referred pre-protocol with an average of 14.2% of these being returned. Post protocol there were 2110 patients referred with 31.2% returned. There was an increase in 195% of referrals returned to the GP (p < 0.0001). The highest proportion of these was for mild to moderate osteoarthritis on the radiograph which has been proven to be unsuitable for intervention. At 12 month analysis there was no significant increase in patients re-referred to the service (p = 0.53) DISCUSSION: The new screening tool allows more appropriate referrals to be seen in clinic allowing less frustration to clinicians and patients by reducing therapeutic inertia. Furthermore it allows new referrals to be seen by the most appropriate sub-specialist. It allows advice to be given to GPs on further management for the patient. 619 appointments were saved. At a cost of £120 per appointment, this leads to a real terms cost saving of £74,280, with further savings in time and travel.

Relapse Prevention and Prediction Strategies in Substance Use Disorder: A Scoping Review.

BACKGROUND: Relapse prevention for those with substance use disorder (SUD) is an evolving practice. Initiatives focused on relapse prevention from other populations may provide the foundation for future considerations and recommendations for recovering anesthesia providers in the workplace. The purpose of this scoping review was to examine what is known about return-to-use prediction and prevention strategies in various populations struggling with SUDs to inform future considerations and implications for recovering anesthesia providers with a history of SUD. METHODS: The Arksey and O'Malley framework was used to conduct a scoping review of the literature. A systematic search was conducted across three databases (PubMed, CINAHL, and PsycInfo) for relevant literature. Search terms used were "measures predicting relapse in substance use disorder" and "relapse prevention in substance use disorder AND anesthesia." Data from articles that met the eligibility criteria were extracted and summarized by the primary author. RESULTS: The search identified 46 articles highlighting various relapse prediction and prevention strategies related to craving and stress, underlying biological factors, neuroimaging, and mindfulness. Relapse prediction and prevention strategies ranged from cell phone applications, monitoring biological markers, and functional neuroimaging of the brain. CONCLUSIONS: Relapse is a concern for individuals with a history of SUD. For anesthesia providers, immediate access to powerful anesthesia medications requires return-to-use prediction and prevention strategies when anesthesia providers return to work after SUD treatment. Although some identified strategies are practical, more research is needed to predict and prevent return to use for recovering anesthesia providers.

An Evaluation of Tertiary Drug Resources' Consistency Regarding Drug-Drug Interactions Between Tricyclic Antidepressants and Herbal Supplements.

Background: Tertiary drug information resources are utilized frequently by health care providers. While pharmacists are uniquely trained and prepared to interpret the information available on these resources, including the results of drug-drug interaction evaluations, discrepancies between such resources pose a major concern for clinicians with regard to patient safety and medication regimen efficacy. It was postulated that drug-drug interaction evaluations between prescription medications and over-the-counter herbal supplements would be particularly problematic. Objective: The objective of this project was to distinguish the discrepancies between tertiary drug information resources in the setting of drug-drug interactions between tricyclic antidepressants (TCAs) and herbal supplements. Methods: The following medications and herbal supplements were evaluated on Lexicomp, Micromedex, and Medscape: amitriptyline, nortriptyline, doxepin, imipramine, desipramine, amoxapine, St. John's Wort, valerian root, ginkgo biloba, and ginseng. Results: While all of the tertiary drug information resources identified a significant reaction between each TCA and St. John's Wort due to the risk of serotonin syndrome, several other discrepancies were noted, with regard to both the severity of the interaction indicated and whether or not an interaction was identified. Conclusion: It is imperative that clinicians be aware of potential discrepancies between tertiary drug information resources, including the potential for variation in both the clinical interpretation of its severity and the recognition of an interaction.

Gene expression reveals immune response strategies of naïve Hawaiian honeycreepers experimentally infected with introduced avian malaria.

The unprecedented rise in the number of new and emerging infectious diseases in the last quarter century poses direct threats to human and wildlife health. The introduction to the Hawaiian archipelago of Plasmodium relictum and the mosquito vector that transmits the parasite has led to dramatic losses in endemic Hawaiian forest bird species. Understanding how mechanisms of disease immunity to avian malaria may evolve is critical as climate change facilitates increased disease transmission to high elevation habitats where malaria transmission has historically been low and the majority of the remaining extant Hawaiian forest bird species now reside. Here, we compare the transcriptomic profiles of highly susceptible Hawai'i 'amakihi (Chlorodrepanis virens) experimentally infected with P. relictum to those of uninfected control birds from a naïve high elevation population. We examined changes in gene expression profiles at different stages of infection to provide an in-depth characterization of the molecular pathways contributing to survival or mortality in these birds. We show that the timing and magnitude of the innate and adaptive immune response differed substantially between individuals that survived and those that succumbed to infection, and likely contributed to the observed variation in survival. These results lay the foundation for developing gene-based conservation strategies for Hawaiian honeycreepers by identifying candidate genes and cellular pathways involved in the pathogen response that correlate with a bird's ability to recover from malaria infection.

Comparative Modeling and Analysis of Extremophilic D-Ala-D-Ala Carboxypeptidases.

Understanding the molecular adaptations of organisms to extreme environments requires a comparative analysis of protein structure, function, and dynamics across species found in different environmental conditions. Computational studies can be particularly useful in this pursuit, allowing exploratory studies of large numbers of proteins under different thermal and chemical conditions that would be infeasible to carry out experimentally. Here, we perform such a study of the MEROPS family S11, S12, and S13 proteases from psychophilic, mesophilic, and thermophilic bacteria. Using a combination of protein structure prediction, atomistic molecular dynamics, and trajectory analysis, we examine both conserved features and trends across thermal groups. Our findings suggest a number of hypotheses for experimental investigation.

Network Hamiltonian Models for Unstructured Protein Aggregates, with Application to γD-Crystallin.

Network Hamiltonian models (NHMs) are a framework for topological coarse-graining of protein-protein interactions, in which each node corresponds to a protein, and edges are drawn between nodes representing proteins that are noncovalently bound. Here, this framework is applied to aggregates of γD-crystallin, a structural protein of the eye lens implicated in cataract disease. The NHMs in this study are generated from atomistic simulations of equilibrium distributions of wild-type and the cataract-causing variant W42R in solution, performed by Wong, E. K.; Prytkova, V.; Freites, J. A.; Butts, C. T.; Tobias, D. J. Molecular Mechanism of Aggregation of the Cataract-Related γD-Crystallin W42R Variant from Multiscale Atomistic Simulations. Biochemistry2019, 58 (35), 3691-3699. Network models are shown to successfully reproduce the aggregate size and structure observed in the atomistic simulation, and provide information about the transient protein-protein interactions therein. The system size is scaled from the original 375 monomers to a system of 10000 monomers, revealing a lowering of the upper tail of the aggregate size distribution of the W42R variant. Extrapolation to higher and lower concentrations is also performed. These results provide an example of the utility of NHMs for coarse-grained simulation of protein systems, as well as their ability to scale to large system sizes and high concentrations, reducing computational costs while retaining topological information about the system.

Mutation Effects on Structure and Dynamics: Adaptive Evolution of the SARS-CoV-2 Main Protease.

The main protease of SARS-CoV-2 (Mpro) plays a critical role in viral replication; although it is relatively conserved, Mpro has nevertheless evolved over the course of the COVID-19 pandemic. Here, we examine phenotypic changes in clinically observed variants of Mpro, relative to the originally reported wild-type enzyme. Using atomistic molecular dynamics simulations, we examine effects of mutation on protein structure and dynamics. In addition to basic structural properties such as variation in surface area and torsion angles, we use protein structure networks and active site networks to evaluate functionally relevant characters related to global cohesion and active site constraint. Substitution analysis shows a continuing trend toward more hydrophobic residues that are dependent on the location of the residue in primary, secondary, tertiary, and quaternary structures. Phylogenetic analysis provides additional evidence for the impact of selective pressure on mutation of Mpro. Overall, these analyses suggest evolutionary adaptation of Mpro toward more hydrophobicity and a less-constrained active site in response to the selective pressures of a novel host environment.

Privacy by Projection: Federated Population Density Estimation by Projecting on Random Features.

We consider the problem of population density estimation based on location data crowdsourced from mobile devices, using kernel density estimation (KDE). In a conventional, centralized setting, KDE requires mobile users to upload their location data to a server, thus raising privacy concerns. Here, we propose a Federated KDE framework for estimating the user population density, which not only keeps location data on the devices but also provides probabilistic privacy guarantees against a malicious server that tries to infer users' location. Our approach Federated random Fourier feature (RFF) KDE leverages a random feature representation of the KDE solution, in which each user's information is irreversibly projected onto a small number of spatially delocalized basis functions, making precise localization impossible while still allowing population density estimation. We evaluate our method on both synthetic and real-world datasets, and we show that it achieves a better utility (estimation performance)-vs-privacy (distance between inferred and true locations) tradeoff, compared to state-of-the-art baselines (e.g., GeoInd). We also vary the number of basis functions per user, to further improve the privacy-utility trade-off, and we provide analytical bounds on localization as a function of areal unit size and kernel bandwidth.

Highly scalable maximum likelihood and conjugate Bayesian inference for ERGMs on graph sets with equivalent vertices.

The exponential family random graph modeling (ERGM) framework provides a highly flexible approach for the statistical analysis of networks (i.e., graphs). As ERGMs with dyadic dependence involve normalizing factors that are extremely costly to compute, practical strategies for ERGMs inference generally employ a variety of approximations or other workarounds. Markov Chain Monte Carlo maximum likelihood (MCMC MLE) provides a powerful tool to approximate the maximum likelihood estimator (MLE) of ERGM parameters, and is generally feasible for typical models on single networks with as many as a few thousand nodes. MCMC-based algorithms for Bayesian analysis are more expensive, and high-quality answers are challenging to obtain on large graphs. For both strategies, extension to the pooled case-in which we observe multiple networks from a common generative process-adds further computational cost, with both time and memory scaling linearly in the number of graphs. This becomes prohibitive for large networks, or cases in which large numbers of graph observations are available. Here, we exploit some basic properties of the discrete exponential families to develop an approach for ERGM inference in the pooled case that (where applicable) allows an arbitrarily large number of graph observations to be fit at no additional computational cost beyond preprocessing the data itself. Moreover, a variant of our approach can also be used to perform Bayesian inference under conjugate priors, again with no additional computational cost in the estimation phase. The latter can be employed either for single graph observations, or for observations from graph sets. As we show, the conjugate prior is easily specified, and is well-suited to applications such as regularization. Simulation studies show that the pooled method leads to estimates with good frequentist properties, and posterior estimates under the conjugate prior are well-behaved. We demonstrate the usefulness of our approach with applications to pooled analysis of brain functional connectivity networks and to replicated x-ray crystal structures of hen egg-white lysozyme.