RESUMO
Predicting infectious disease dynamics is a central challenge in disease ecology. Models that can assess which individuals are most at risk of being exposed to a pathogen not only provide valuable insights into disease transmission and dynamics but can also guide management interventions. Constructing such models for wild animal populations, however, is particularly challenging; often only serological data are available on a subset of individuals and nonlinear relationships between variables are common. Here we provide a guide to the latest advances in statistical machine learning to construct pathogen-risk models that automatically incorporate complex nonlinear relationships with minimal statistical assumptions from ecological data with missing data. Our approach compares multiple machine learning algorithms in a unified environment to find the model with the best predictive performance and uses game theory to better interpret results. We apply this framework on two major pathogens that infect African lions: canine distemper virus (CDV) and feline parvovirus. Our modelling approach provided enhanced predictive performance compared to more traditional approaches, as well as new insights into disease risks in a wild population. We were able to efficiently capture and visualize strong nonlinear patterns, as well as model complex interactions between variables in shaping exposure risk from CDV and feline parvovirus. For example, we found that lions were more likely to be exposed to CDV at a young age but only in low rainfall years. When combined with our data calibration approach, our framework helped us to answer questions about risk of pathogen exposure that are difficult to address with previous methods. Our framework not only has the potential to aid in predicting disease risk in animal populations, but also can be used to build robust predictive models suitable for other ecological applications such as modelling species distribution or diversity patterns.
Assuntos
Vírus da Cinomose Canina , Leões , Animais , Animais Selvagens , Ecologia , Aprendizado de MáquinaRESUMO
BACKGROUND: Alcohol use disorder (AUD) has been described as a chronic disease given the high rates that affected individuals have in returning to drinking after a change attempt. Many studies have characterized predictors of aggregated alcohol use (e.g., percent heavy drinking days) following treatment for AUD. However, to inform future research on predicting drinking as an AUD outcome measure, a better understanding is needed of the patterns of drinking that surround a treatment episode and which clinical measures predict patterns of drinking. METHODS: We analyzed data from the Project MATCH and COMBINE studies (MATCH: n = 1726; 24.3% female, 20.0% non-White; COMBINE: n = 1383; 30.9% female, 23.2% non-White). Daily drinking was measured in the 90 days prior to treatment, 90 days (MATCH) and 120 days (COMBINE) during treatment, and 365 days following treatment. Gradient boosting machine learning methods were used to explore baseline predictors of drinking patterns. RESULTS: Drinking patterns during a prior time period were the most consistent predictors of future drinking patterns. Social network drinking, AUD severity, mental health symptoms, and constructs based on the addiction cycle (incentive salience, negative emotionality, and executive function) were associated with patterns of drinking prior to treatment. Addiction cycle constructs, AUD severity, purpose in life, social network, legal history, craving, and motivation were associated with drinking during the treatment period and following treatment. CONCLUSIONS: There is heterogeneity in drinking patterns around an AUD treatment episode. This study provides novel information about variables that may be important to measure to improve the prediction of drinking patterns during and following treatment. Future research should consider which patterns of drinking they aim to predict and which period of drinking is most important to predict. The current findings could guide the selection of predictor variables and generate hypotheses for those predictors.
RESUMO
To date, PHMR has often relied on male/female stratification, but rarely considers the complex, intersecting social positions of men and women in describing the prevalence of health and disease. Stratification on an Intersectional Gender-Score (IG-Score), which is based on a variety of social covariables, would allow comparison of the prevalence of individuals who share the same complex intersectional profile (IG-Score). The cross-sectional case study was based on the German Socio-Economic Panel 2017 (n = 23,269 age 18+). After stratification, covariable-balance within the total sample and IG-Score-subgroups was assessed by standardized mean differences. Prevalence of self-rated health, mental distress, depression and hypertension was compared in men and women. In the IG-Score-subgroup with highest proportion of males and lowest probability of falling into the 'woman'-category, most individuals were in full-time employment. The IG-Score-subgroup with highest proportion of women and highest probability of falling into the 'woman'-category was characterized by part-time/occasional employment, housewife/-husband, and maternity/parental leave. Gender differences in prevalence of health indicators remained within the male-dominated IG-Score-subgroup, whereas the same prevalence of depression and self-rated health was observed for men and women constituting the female-dominated IG-Score-subgroup. These results might indicate that sex/gender differences of depression and self-rated health could be interpreted against the background of gender associated processes. In summary, the proposed procedure allows comparison of prevalence of health indicators conditional on men and women sharing the same complex intersectional profile.
Assuntos
Enquadramento Interseccional , Saúde Pública , Gravidez , Humanos , Masculino , Feminino , Adolescente , Estudos Transversais , Fatores Sexuais , EmpregoRESUMO
We introduce a new R package "MrIML" ("Mister iml"; Multi-response Interpretable Machine Learning). MrIML provides a powerful and interpretable framework that enables users to harness recent advances in machine learning to quantify multilocus genomic relationships, to identify loci of interest for future landscape genetics studies, and to gain new insights into adaptation across environmental gradients. Relationships between genetic variation and environment are often nonlinear and interactive; these characteristics have been challenging to address using traditional landscape genetic approaches. Our package helps capture this complexity and offers functions that fit and interpret a wide range of highly flexible models that are routinely used for single-locus landscape genetics studies but are rarely extended to estimate response functions for multiple loci. To demonstrate the package's broad functionality, we test its ability to recover landscape relationships from simulated genomic data. We also apply the package to two empirical case studies. In the first, we model genetic variation of North American balsam poplar (Populus balsamifera, Salicaceae) populations across environmental gradients. In the second case study, we recover the landscape and host drivers of feline immunodeficiency virus genetic variation in bobcats (Lynx rufus). The ability to model thousands of loci collectively and compare models from linear regression to extreme gradient boosting, within the same analytical framework, has the potential to be transformative. The MrIML framework is also extendable and not limited to modelling genetic variation; for example, it can quantify the environmental drivers of microbiomes and coinfection dynamics.