Interpreting generative adversarial networks to infer natural selection from genetic data.

Understanding natural selection and other forms of non-neutrality is a major focus for the use of machine learning in population genetics. Existing methods rely on computationally intensive simulated training data. Unlike efficient neutral coalescent simulations for demographic inference, realistic simulations of selection typically require slow forward simulations. Because there are many possible modes of selection, a high dimensional parameter space must be explored, with no guarantee that the simulated models are close to the real processes. Finally, it is difficult to interpret trained neural networks, leading to a lack of understanding about what features contribute to classification. Here we develop a new approach to detect selection and other local evolutionary processes that requires relatively few selection simulations during training. We build upon a generative adversarial network trained to simulate realistic neutral data. This consists of a generator (fitted demographic model), and a discriminator (convolutional neural network) that predicts whether a genomic region is real or fake. As the generator can only generate data under neutral demographic processes, regions of real data that the discriminator recognizes as having a high probability of being "real" do not fit the neutral demographic model and are therefore candidates for targets of selection. To incentivize identification of a specific mode of selection, we fine-tune the discriminator with a small number of custom non-neutral simulations. We show that this approach has high power to detect various forms of selection in simulations, and that it finds regions under positive selection identified by state-of-the-art population genetic methods in three human populations. Finally, we show how to interpret the trained networks by clustering hidden units of the discriminator based on their correlation patterns with known summary statistics.

INTERPRETING GENERATIVE ADVERSARIAL NETWORKS TO INFER NATURAL SELECTION FROM GENETIC DATA.

Understanding natural selection in humans and other species is a major focus for the use of machine learning in population genetics. Existing methods rely on computationally intensive simulated training data. Unlike efficient neutral coalescent simulations for demographic inference, realistic simulations of selection typically requires slow forward simulations. Because there are many possible modes of selection, a high dimensional parameter space must be explored, with no guarantee that the simulated models are close to the real processes. Mismatches between simulated training data and real test data can lead to incorrect inference. Finally, it is difficult to interpret trained neural networks, leading to a lack of understanding about what features contribute to classification. Here we develop a new approach to detect selection that requires relatively few selection simulations during training. We use a Generative Adversarial Network (GAN) trained to simulate realistic neutral data. The resulting GAN consists of a generator (fitted demographic model) and a discriminator (convolutional neural network). For a genomic region, the discriminator predicts whether it is "real" or "fake" in the sense that it could have been simulated by the generator. As the "real" training data includes regions that experienced selection and the generator cannot produce such regions, regions with a high probability of being real are likely to have experienced selection. To further incentivize this behavior, we "fine-tune" the discriminator with a small number of selection simulations. We show that this approach has high power to detect selection in simulations, and that it finds regions under selection identified by state-of-the art population genetic methods in three human populations. Finally, we show how to interpret the trained networks by clustering hidden units of the discriminator based on their correlation patterns with known summary statistics. In summary, our approach is a novel, efficient, and powerful way to use machine learning to detect natural selection.

Harris County Public Health Mosquito and Vector Control Division Emergency Response to Hurricane Harvey: Vector-Borne Disease Surveillance and Control.

Hurricane Harvey made a landfall on the Texas Gulf Coast on August 25, 2017, stalling over Harris County as a tropical storm for 4 days (August 26-29), dumping approximately 127 cm of rain. This tremendous amount of rainfall overwhelmed the county's natural and man-made drainage systems, resulting in unprecedented widespread flooding. Immediately following, Harris County Public Health Mosquito and Vector Control Division conducted a countywide emergency vector control response by integrating surveillance, control, and education strategies. This included landing rate counts, mosquito and avian surveillance, arbovirus testing, ground-based ultra-low volume (ULV) and aerial pesticide spraying, and community outreach. The immediate response lasted for 4 wk through September, resulting in 774 landing rates, 49,342 ha treated by ground-based ULV, 242,811 ha treated by aerial ULV, 83,241 mosquitoes collected, 1,807 mosquito pools tested, and 20 education/outreach sessions. Recovery activities of 3 additional education/outreach events continued through October while surveillance and control activities returned to routine status.

Tumor re-growth, case outcome, and tumor scoring systems in rehabilitated green turtles with fibropapillomatosis.

Fibropapillomatosis (FP) is an infectious, neoplastic disease of major concern in sea turtle rehabilitation facilities. Rehabilitating sea turtles that undergo tumor removal surgery often have tumor regrowth and may experience mortality. We evaluated tumor score, removal, and regrowth in rehabilitating green sea turtles with FP in 4 rehabilitation facilities in the southeastern USA during 2009-2017. Of 756 cases, 312 (41%) underwent tumor removal surgery, 155 (50%) of those had tumor regrowth within an average of 46 ± 45 d, and 85 (27%) had multiple (>1) regrowth events. Of 756 turtles with FP, 563 (75%) did not survive after admission into a rehabilitation facility, including 283 (37%) that were euthanized and 280 that died without euthanasia (37%), and 193 survived, including 186 (25%) released and 7 (1%) placed in permanent captive care. Tumor removal surgery increased the odds of tumor regrowth but also enhanced survivorship, whereas tumor regrowth was not a significant predictor of case outcome. Three FP tumor scoring systems were used to assign tumor scores to 449 cases, and differing results emphasize that tumor scoring systems should be applied to the situations and/or location(s) for which they were intended. FP tumor score was not a significant predictor for the event or extent of FP tumor regrowth after surgical excision. Under current rehabilitation regimes, outcomes of rehabilitation for tumored turtles have a low probability of success. The results of this study may be used to help guide clinical decision-making and determine prognoses for rehabilitating sea turtles with FP.

CULTURAL ADAPTATIONS OF EVIDENCE-BASED HOME-VISITATION MODELS IN TRIBAL COMMUNITIES.

The Tribal Maternal, Infant, and Early Childhood Home Visiting (Tribal MIECHV) Program provides federal grants to tribes, tribal consortia, tribal organizations, and urban Indian organizations to implement evidence-based home-visiting services for American Indian and Alaska Native (AI/AN) families. To date, only one evidence-based home-visiting program has been developed for use in AI/AN communities. The purpose of this article is to describe the steps that four Tribal MIECHV Programs took to assess community needs, select a home-visiting model, and culturally adapt the model for use in AI/AN communities. In these four unique Tribal MIECHV Program settings, each program employed a rigorous needs-assessment process and developed cultural modifications in accordance with community strengths and needs. Adaptations occurred in consultation with model developers, with consideration of the conceptual rationale for the program, while grounding new content in indigenous cultures. Research is needed to improve measurement of home-visiting outcomes in tribal and urban AI/AN settings, develop culturally grounded home-visiting interventions, and assess the effectiveness of home visiting in AI/AN communities.

Are there too few women presenting at emergency medicine conferences?

INTRODUCTION: There is a perception that women are under-represented as speakers at emergency medicine (EM) conferences. We aimed to evaluate the ratio of male to female speakers and the proportion of presenting time by gender at major international EM conferences. METHODS: Conference programmes of the major English-speaking EM conferences occurring from 2014 to 2015 were obtained. The number of presentations, the gender of the speaker and the duration of each presentation were recorded. RESULTS: We analysed eight major EM conferences. These included 2382 presentations, of which 29.9% (range 22.5%-40.9%) were given by women. In total, 56 104 min of presentations were analysed, of which 27.6% (range 21%-36.7%) were delivered by women. On average, presentations by women were 95 s shorter than presentations by men (23 vs 21 min 25 s). CONCLUSIONS: Male speakers exceed female speakers at major EM conferences. The reasons for this imbalance are likely complex and multifactorial and may reflect the gender imbalance within the specialty.

Perceived stigma in persons with early-stage dementia: Longitudinal findings: Part 1.

This longitudinal study examined perceived stigma in persons with dementia, with 50 persons with dementia, and 47 corresponding family caregivers. Data were collected at baseline and at 6, 12, and 18 months. Study results are reported in two parts, with findings regarding the stability of perceived stigma, measured using the modified Stigma Impact Scale, and relationship of stigma to person-centered variables being reported here. Findings included stability in perceived stigma, which did not show a downward trend until 18 months. Significant differences at baseline were found only for geographic location (rural vs. urban) with persons living in urban areas having higher levels of Stigma Impact Scale internalized shame compared to rural counterparts. Cognitive functioning was significantly, positively related to the Stigma Impact Scale social rejection and social isolation subscales. Findings support the enduring nature of perceived stigma over the early disease stages and the relationship of perceived stigma to some person-centered characteristics.

Anxiety and stigma in dementia: a threat to aging in place.

The number of Americans with dementia is expected to increase as the population ages. Developing dementia is feared by many older adults and may result in anxiety in persons with dementia. This article focuses on anxiety, one of the least understood symptoms associated with dementia in community-dwelling older adults, the stigma of dementia, and the relationship between anxiety and stigma in dementia. When undetected and untreated, anxiety and associated stigma can adversely affect quality of life and the ability to age in place.

Coronary artery endothelial transcriptome in vivo: identification of endoplasmic reticulum stress and enhanced reactive oxygen species by gene connectivity network analysis.

BACKGROUND: Endothelial function is central to the localization of atherosclerosis. The in vivo endothelial phenotypic footprints of arterial bed identity and site-specific atherosusceptibility are addressed. METHODS AND RESULTS: Ninety-eight endothelial cell samples from 13 discrete coronary and noncoronary arterial regions of varying susceptibilities to atherosclerosis were isolated from 76 normal swine. Transcript profiles were analyzed to determine the steady-state in vivo endothelial phenotypes. An unsupervised systems biology approach using weighted gene coexpression networks showed highly correlated endothelial genes. Connectivity network analysis identified 19 gene modules, 12 of which showed significant association with circulatory bed classification. Differential expression of 1300 genes between coronary and noncoronary artery endothelium suggested distinct coronary endothelial phenotypes, with highest significance expressed in gene modules enriched for biological functions related to endoplasmic reticulum (ER) stress and unfolded protein binding, regulation of transcription and translation, and redox homeostasis. Furthermore, within coronary arteries, comparison of endothelial transcript profiles of susceptible proximal regions to protected distal regions suggested the presence of ER stress conditions in susceptible sites. Accumulation of reactive oxygen species throughout coronary endothelium was greater than in noncoronary endothelium consistent with coronary artery ER stress and lower endothelial expression of antioxidant genes in coronary arteries. CONCLUSIONS: Gene connectivity analyses discriminated between coronary and noncoronary endothelial transcript profiles and identified differential transcript levels associated with increased ER and oxidative stress in coronary arteries consistent with enhanced susceptibility to atherosclerosis.

Prelesional arterial endothelial phenotypes in hypercholesterolemia: universal ABCA1 upregulation contrasts with region-specific gene expression in vivo.

Atherosclerosis originates as focal arterial lesions having a predictable distribution to regions of bifurcations, branches, and inner curvatures where blood flow characteristics are complex. Distinct endothelial phenotypes correlate with regional hemodynamics. We propose that systemic risk factors modify regional endothelial phenotype to influence focal susceptibility to atherosclerosis. Transcript profiles of freshly isolated endothelial cells from three atherosusceptible and three atheroprotected arterial regions in adult swine were analyzed to determine the initial prelesional effects of hypercholesterolemia on endothelial phenotypes in vivo. Cholesterol efflux transporter ATP-binding cassette transporter A1 (ABCA1) was upregulated at all sites in response to short-term high-fat diet. Proinflammatory and antioxidative endothelial gene expression profiles were induced in atherosusceptible and atheroprotected regions, respectively. However, markers for endoplasmic reticulum stress, a signature of susceptible endothelial phenotype, were not further enhanced by brief hypercholesterolemia. Both region-specific and ubiquitous (ABCA1) phenotype changes were identified as early prelesional responses of the endothelium to hypercholesterolemia.

Chronic endoplasmic reticulum stress activates unfolded protein response in arterial endothelium in regions of susceptibility to atherosclerosis.

RATIONALE: Endothelial function and dysfunction are central to the focal origin and regional development of atherosclerosis; however, an in vivo endothelial phenotypic footprint of susceptibility to atherosclerosis preceding pathological change remains elusive. OBJECTIVE: To conduct a comparative multi-site genomics study of arterial endothelial phenotype in atherosusceptible and atheroprotected regions. METHODS AND RESULTS: Transcript profiles of freshly isolated endothelial cells from 7 discrete arterial regions in normal swine were analyzed to determine the steady state in vivo endothelial phenotypes in regions of varying susceptibilities to atherosclerosis. The most abundant common feature of the endothelium of all atherosusceptible regions was the upregulation of genes associated with endoplasmic reticulum (ER) stress. The unfolded protein response pathway, induced by ER stress, was therefore investigated in detail in endothelium of the atherosusceptible aortic arch and was found to be partially activated. ER transmembrane signal transducers IRE1alpha and ATF6alpha and their downstream effectors, but not PERK, were activated concomitant with a higher transcript expression of protein folding enzymes and chaperones, indicative of ER stress in vivo. CONCLUSIONS: The findings demonstrate the prevalence of chronic endothelial ER stress and activated unfolded protein response in vivo at atherosusceptible arterial sites. We propose that chronic localized biological stress is linked to spatial susceptibility of the endothelium to the initiation of atherosclerosis.

Contrasting selection pressures on components of the Ras-mediated signal transduction pathway in Drosophila.

The molecular genetics of several signal transduction pathways is well characterized, providing an opportunity to address the nature of the population genetic forces acting on functionally related suites of pleiotropic regulatory genes. Signal transduction is the process by which signals are transmitted from the cell surface to the nucleus or other cellular structures. It plays a fundamental role in regulating a wide range of developmental and physiological processes, many of which are likely to be subject to buffering mechanisms. Here we infer that contrasting selection pressures act on six components of the Ras signal transduction pathway by comparing sequences obtained from 25 alleles of Drosophila melanogaster with one allele of the sibling species D. simulans. The three most upstream components of the cascade, Ras, Drk and polehole, experience strong purifying selection, as they show no fixed amino acid differences between the species and just a handful of rare replacement polymorphisms within D. melanogaster. This portion of the pathway is likely to act as a control point in signal transduction, because the more downstream components Dsor1, corkscrew and Ksr, each show several amino acid replacements between the species. Furthermore, Ksr is nearly monomorphic within D. melanogaster, and application of the HKA and McDonald and Kreitman tests indicate that this gene may have experienced a recent selective sweep, suggesting that modifiers of Ras kinase signalling are the most likely source of quantitative variation associated with this core regulatory pathway.

Cutting edge: inhibiting measles virus infection but promoting reproduction: an explanation for splicing and tissue-specific expression of CD46.

Membrane cofactor protein (MCP; CD46) regulates the complement cascade by inhibiting C3b and C4b deposited on self tissue. This function resides in the complement control protein repeats (CCPs), with CCPs 2-4 essential for regulation. MCP is expressed on the inner acrosomal membrane of human sperm, and Abs to CCP1 inhibit sperm-egg interactions. In somatic tissues, New World monkeys express an alternatively spliced form of MCP lacking CCP1. Although retaining complement-regulatory activity, this form is postulated to render these species less susceptible to strains of the measles virus whose hemagglutinin requires CCP1 and CCP2 for attachment. Using PCR, sequencing, Western blotting, and immunohistochemistry, we characterized MCP expression in the testes and sperm of two New World monkeys. In these species, sperm express MCP bearing CCP1. The germ cell-specific expression pattern of this domain strongly suggests an evolutionarily conserved role for MCP in fertilization.

Characterization of human membrane cofactor protein (MCP; CD46) on spermatozoa.

Membrane cofactor protein (MCP; CD46) is a complement regulator widely expressed as four isoforms that arise via alternative splicing. On human spermatozoa, MCP is expressed on the inner acrosomal membrane and alterations of spermatozoa MCP may be associated with infertility. In rodents, expression of MCP is largely restricted to the testes. MCP on human spermatozoa has a unique M(r) pattern that we have investigated. We also characterized MCP expression in mice transgenic (tg) for human MCP. Human MCP expression in the tg mice mimics the human pattern in that it is located on the inner acrosomal membrane and has a faster M(r) than MCP expressed elsewhere. Sequencing of RT-PCR products from the testis indicates that there is not a unique male reproductive tissue specific cytoplasmic tail. Instead, human spermatozoa express MCP bearing cytoplasmic tail two, which is also utilized in most other tissues and contains several signaling motifs. Further, using N-glycosidases, we demonstrate that the unique lower molecular weight of MCP on spermatozoa is secondary to a modification in the N-linked sugars. Specifically, as the spermatozoa mature, but before they reach the epididymis, the three N-linked sugars of MCP are trimmed to less complex structures. While the purpose of this deglycosylation is unknown, we propose that it is a common feature of proteins expressed on the plasma and inner acrosomal membranes of spermatozoa and hypothesize that it is a spermatozoa specific event critical for facilitating sperm-egg interactions.

Outcomes and secondary prevention strategies for male hip fractures.

OBJECTIVE: To assess clinical outcomes and determine whether osteoporosis assessment and secondary prevention strategies were performed for male veterans hospitalized for hip fractures. DESIGN: Retrospective chart review for male veterans hospitalized for hip fracture from January 1993 through July 1999. SETTING: The Veterans Affairs Medical Center, Madison, WI. RESULTS: Medical charts were available for 46 of 53 male patients admitted for hip fracture during the study period. Three subjects were excluded because hip fracture was associated with high-impact trauma. Mean age of the 43 study patients was 72 years (range 43-91 y), and mean length of hospitalization was 16 days (median 11 d, range 3-108 d). Thirty-two (82%) of 39 veterans whose disposition was documented were discharged to a nursing home. Eleven (26%) of 43 men died within 12 months after fracture. Twelve (28%) had fractured previously. Four (10%) subsequently had another fracture. Three of 9 patients with documented ambulation status were ambulatory at 1 year. Three patients received a bone mass measurement within a prespecified time interval of 6 months subsequent to fracture. No patient's records included a diagnosis of osteoporosis either before or within 6 months after fracture. One-third of the patients had documentation of calcium or multivitamin supplementation at discharge. One patient was receiving calcitonin at the time of fracture and continued to receive it afterward. No other patient was prescribed antiresorptive therapy by the time of hospital discharge. CONCLUSIONS: Male veterans with hip fractures received inadequate evaluation and treatment for osteoporosis, although a substantial portion had documentation of recurrent fractures. Education of clinicians and creation of algorithms for management of established osteoporosis may improve outcomes for these individuals.