로지스틱회귀모형의 평가를 위한 그래픽적 방법

Computer simulations are routinely conducted to evaluate new statistical methods, to compare the properties among different methods, and to mimic the observed data in genetic epidemiology studies. Conducting simulation studies can become a complicated task as several challenges can occur, such as the selection of an appropriate simulation tool and the specification of parameters in the simulation model. Although abundant simulated data have been generated for human genetic research, currently there is no public database designed specifically as a repository for these simulated data. With the lack of such a database, for similar studies, similar simulations may have been repeated, which resulted in redundant work. Thus, we created an online platform, the Genetic Epidemiology Simulation Database (GESDB), for simulation data sharing and discussion of simulation techniques for genetic epidemiology studies. GESDB consists of a database for storing simulation scripts, simulated data and documentation from published articles as well as a discussion forum, which provides a platform for discussion of the simulated data and exchanging simulation ideas. Moreover, summary statistics such as the simulation tools that are most commonly used and datasets that are most frequently downloaded are provided. The statistics will be informative for researchers to choose an appropriate simulation tool or select a common dataset for method comparisons. GESDB can be accessed at http://gesdb.nhri.org.tw.Database URL:http://gesdb.nhri.org.tw

BackgroundColocalization is a statistical method used in genetics to determine whether the same variant is causal for multiple phenotypes, for example, complex traits and gene expression. It provides stronger mechanistic evidence than shared significance, which can be produced through separate causal variants in linkage disequilibrium. Current colocalization methods require full summary statistics for both traits, limiting their use with the majority of reported GWAS associations (e.g. GWAS Catalog). We propose a new approximation to the popular coloc method that can be applied when limited summary statistics are available.Our method (POint EstiMation of Colocalization, POEMColoc) imputes missing summary statistics for one or both traits using LD structure in a reference panel, and performs colocalization using the imputed summary statistics.ResultsWe evaluate the performance of POEMColoc using real (UK Biobank phenotypes and GTEx eQTL) and simulated datasets. We show good correlation between posterior probabilities of colocalization computed from imputed and observed datasets and similar accuracy in simulation. We evaluate scenarios that might reduce performance and show that multiple independent causal variants in a region and imputation from a limited subset of typed variants have a larger effect while mismatched ancestry in the reference panel has a modest effect. Further, we find that POEMColoc is a better approximation of coloc when the imputed association statistics are from a well powered study (e.g., relatively larger sample size or effect size). Applying POEMColoc to estimate colocalization of GWAS Catalog entries and GTEx eQTL, we find evidence for colocalization of 150,000 trait-gene-tissue triplets.ConclusionsWe find that colocalization analysis performed with full summary statistics can be closely approximated when only the summary statistics of the top SNP are available for one or both traits. When applied to the full GWAS Catalog and GTEx eQTL, we find that colocalized trait-gene pairs are enriched in tissues relevant to disease etiology and for matches to approved drug mechanisms. POEMColoc R package is available at https://github.com/AbbVie-ComputationalGenomics/POEMColoc.

This paper describes the results of exploratory analyses of black box simulation data modeling crowds exiting different configurations of a one-story building. The simulation data was created using the SteerSuite platform. Exploratory analysis was performed on the simulation data without knowledge of simulation algorithm. The analysis effort provided a hands-on introduction to issues in crowd dynamics. Analyses focused on visualization, panic detection, exit convergence pattern discovery, identification of parameters influencing exit times, and estimation of exit times. A variety of mathematical and statistical methods were used: k-means clustering, principal component analysis, normalized cut grouping, product formula representation of dyadic measures, logistic regression, auto-encoders, and neural networks. The combined set of results provided insight into the algorithm and the behavior modeled by the algorithm and revealed the need for quantitative features modeling and distinguishing the shapes of the building configurations.

Approximate Bayesian computation (ABC) has become one of the major tools of likelihood-free statistical inference in complex mathematical models. Simultaneously, stochastic differential equations (SDEs) have developed to an established tool for modelling time-dependent, real-world phenomena with underlying random effects. When applying ABC to stochastic models, two major difficulties arise: First, the derivation of effective summary statistics and proper distances is particularly challenging, since simulations from the stochastic process under the same parameter configuration result in different trajectories. Second, exact simulation schemes to generate trajectories from the stochastic model are rarely available, requiring the derivation of suitable numerical methods for the synthetic data generation. To obtain summaries that are less sensitive to the intrinsic stochasticity of the model, we propose to build up the statistical method (e.g. the choice of the summary statistics) on the underlying structural properties of the model. Here, we focus on the existence of an invariant measure and we map the data to their estimated invariant density and invariant spectral density. Then, to ensure that these model properties are kept in the synthetic data generation, we adopt measure-preserving numerical splitting schemes. The derived property-based and measure-preserving ABC method is illustrated on the broad class of partially observed Hamiltonian type SDEs, both with simulated data and with real electroencephalography data. The derived summaries are particularly robust to the model simulation, and this fact, combined with the proposed reliable numerical scheme, yields accurate ABC inference. In contrast, the inference returned using standard numerical methods (Euler–Maruyama discretisation) fails. The proposed ingredients can be incorporated into any type of ABC algorithm and directly applied to all SDEs that are characterised by an invariant distribution and for which a measure-preserving numerical method can be derived.

Based on benchmarking various methods and analyzing multiple real scRNA-seq datasets, a computational platform/workflow and a set of tips for best practices are developed for analyzing causal interactions or relationships in single cells.

Exploration of PRO-CTCAE Data Used for Exposure-Response Relationships in an Oncology Clinical Trial

Objective The primary purpose of this study is to assess the variance in the effects of factors influencing ever migrated Malawian population’s subjective HIV infection expectation. Design Using data from the Malawi Longitudinal Study of Families and Health (MLSFH) survey (1998-2010), 7805 ever migrated Malawian adults were selected for the study. Summary statistics, logistic regression and longitudinal multi-level models were fitted for the study. A binary logistic regression was used to estimate the direction and magnitude of the associations between the variables selected for the study Five multilevel models with random intercepts and coefficients nominal response were fitted. Results The study revealed that sexual behaviours had the most significant effect on ever migrated Malawian’s subjective HIV infection expectation. All metrics showed that the conditional growth model had the most signficant outcome. The addition of time and other variables as predictors had a significant effect on the conditional growth model. Conclusion Interventions designed to decrease the spread of HIV should target sexual behaviours and widespread testing among the ever migrated population to reduce subjective HIV infection thoughts.

Graphical approach provides a useful framework for multiplicity adjustment in clinical trials with multiple endpoints. When designing a graphical approach, initial weight and transition probability for the endpoints are often assigned based on clinical importance. For example, practitioners may prefer putting more weights on some primary endpoints. The clinical preference can be formulated as a constrain in the sample size optimization problem. However, there has been a lack of theoretical guidance on how to specify initial weight and transition probability in a graphical approach to meet the clinical preference but at the same time to minimize the sample size needed for a power requirement. To fill this gap, we propose statistical methods to optimize sample size over initial weight and transition probability in a graphical approach under a common setting, which is to use marginal power for each endpoint in a trial design. Importantly, we prove that some of the commonly used graphical approaches such as putting all initial weights on one endpoint are suboptimal. Our methods are flexible, which can be used for both single-arm trials and randomized controlled trials with either continuous or binary or mixed types of endpoints. Additionally, we prove the existence of optimal solution where all marginal powers are placed exactly at the prespecified values, assuming continuity. Two hypothetical clinical trial designs are presented to illustrate the application of our methods under different scenarios. Results are first presented for a design with two endpoints and are further generalized to three or more endpoints. Our findings are helpful to guide the design of a graphical approach and the sample size calculation in clinical trials.

The advent of large biobanks has substantially increased the accuracy of polygenic scores (PGS). However, most existing PGSs were derived from European-ancestry data and often exhibit reduced predictive performance when applied to individuals of non-European ancestries. Transfer Learning offers a promising strategy to address this limitation by leveraging information learned in one population to improve prediction in another. Here, we introduce GPTL, an R package that implements three Transfer Learning based approaches for developing PGS: (1) gradient descent with early stopping, (2) a penalized regression model that shrinks variant-effect estimates toward prior values, and (3) a Bayesian method with a finite-mixture prior that enables integration of multiple prior sources of information. Using both simulated data and real data from the UK-Biobank and All of Us, we demonstrate that PGS generated with GPTL's Transfer Learning algorithms consistently outperform single-ancestry PGS and, in many settings, match or exceed the performance of multi-ancestry ensemble-based PGS. Our software can be used with either individual genotype-phenotype data or summary statistics from genome-wide association studies.

There is growing interest in testing genetic pleiotropy, which is when a single genetic variant influences multiple traits. Several methods have been proposed; however, these methods have some limitations. First, all the proposed methods are based on the use of individual-level genotype and phenotype data; in contrast, for logistical, and other, reasons, summary statistics of univariate SNP-trait associations are typically only available based on meta- or mega-analyzed large genome-wide association study (GWAS) data. Second, existing tests are based on marginal pleiotropy, which cannot distinguish between direct and indirect associations of a single genetic variant with multiple traits due to correlations among the traits. Hence, it is useful to consider conditional analysis, in which a subset of traits is adjusted for another subset of traits. For example, in spite of substantial lowering of low-density lipoprotein cholesterol (LDL) with statin therapy, some patients still maintain high residual cardiovascular risk, and, for these patients, it might be helpful to reduce their triglyceride (TG) level. For this purpose, in order to identify new therapeutic targets, it would be useful to identify genetic variants with pleiotropic effects on LDL and TG after adjusting the latter for LDL; otherwise, a pleiotropic effect of a genetic variant detected by a marginal model could simply be due to its association with LDL only, given the well-known correlation between the two types of lipids. Here, we develop a new pleiotropy testing procedure based only on GWAS summary statistics that can be applied for both marginal analysis and conditional analysis. Although the main technical development is based on published union-intersection testing methods, care is needed in specifying conditional models to avoid invalid statistical estimation and inference. In addition to the previously used likelihood ratio test, we also propose using generalized estimating equations under the working independence model for robust inference. We provide numerical examples based on both simulated and real data, including two large lipid GWAS summary association datasets based on ∼100,000 and ∼189,000 samples, respectively, to demonstrate the difference between marginal and conditional analyses, as well as the effectiveness of our new approach.

This paper compares three published methods for analysing multiple correlated ROC curves: a method using generalized estimating equations with marginal non-proportional ordinal regression models; a method using jackknifed pseudovalues of summary statistics; a method using a corrected F-test from analysis of variance of summary statistics. Use of these methods is illustrated through six real data examples from studies with the common factorial design, that is, multiple readers interpreting images obtained with each test modality on each study subject. The issue of the difference between typical summary statistics and summary statistics from typical ROC curves is explored. The examples also address similarities and differences among the analytical methods. In particular, while point estimates of differences between test modalities are similar, the standard errors of these differences do not agree for all three methods. A simulation study supports the standard errors provided by the generalized estimating equations with marginal non-proportional ordinal regression models.

Detecting Discrimination: Analyzing Racial Disparities in Public Contracting

Biological invasions reshape environments and affect the ecological and economic welfare of states and communities. Such invasions advance on multiple spatial scales, complicating their control. When modeling stochastic dispersal processes, intractable likelihoods and autocorrelated data complicate parameter estimation. As with other approaches, the recent synthetic likelihood framework for stochastic models uses summary statistics to reduce this complexity; however, it additionally provides usable likelihoods, facilitating the use of existing likelihood-based machinery. Here, we extend this framework to parameterize multi-scale spatio-temporal dispersal models and compare existing and newly developed spatial summary statistics to characterize dispersal patterns. We provide general methods to evaluate potential summary statistics and present a fitting procedure that accurately estimates dispersal parameters on simulated data. Finally, we apply our methods to quantify the short and long range dispersal of Chagas disease vectors in urban Arequipa, Peru, and assess the feasibility of a purely reactive strategy to contain the invasion.

Reconstructing the evolutionary history of great apes is of particular importance for our understanding of the demographic history of humans. The reason for this is that modern humans and their hominin ancestors evolved in Africa and thus shared the continent with the ancestors of chimpanzees and gorillas. Common chimpanzees (Pan troglodytes) are our closest relatives with bonobos (Pan paniscus) and most of what we know about their evolutionary history comes from genetic and genomic studies. Most evolutionary studies of common chimpanzees have assumed that the four currently recognised subspecies can be modelled using simple tree models where each subspecies is panmictic and represented by one branch of the evolutionary tree. In addition, one recent genetic study claimed to have detected admixture between bonobos and chimpanzees. However, several studies have identified the existence of significant population structure with evidence of isolation-by-distance (IBD) patterns, both within and between subspecies. This suggests that demographic models integrating population structure may be necessary to improve our understanding of their evolutionary history. Here we propose to use n-island models within each subspecies to infer a demographic history integrating population structure and changes in connectivity (i.e., gene flow). For each subspecies, we use SNIF (Structured non-stationary inference framework), a method developed to infer a piecewise stationary n-island model using PSMC (Pairwise sequentially Markovian coalescent) curves as summary statistics. We then propose a general model integrating the four subspecies as metapopulations within a phylogenetic tree. We find that this model correctly predicts estimates of within subspecies genetic diversity and differentiation, but overestimates genetic differentiation between subspecies as a consequence of the tree structure. We argue that spatial models integrating gene flow between subspecies should improve the prediction of between subspecies differentiation and generate the observed IBD patterns. We also simulated data under a simple spatially structured model for bonobos and chimpanzees (without admixture) and found that it generates potentially spurious signals of admixture between the two species that have been reported and could thus be spurious. This may have implications for our understanding of the evolutionary history of the Homo genus.

Increasing attention has been devoted to estimation of species-level phylogenetic relationships under the coalescent model. However, existing methods either use summary statistics (gene trees) to carry out estimation, ignoring an important source of variability in the estimates, or involve computationally intensive Bayesian Markov chain Monte Carlo algorithms that do not scale well to whole-genome datasets. We develop a method to infer relationships among quartets of taxa under the coalescent model using techniques from algebraic statistics. Uncertainty in the estimated relationships is quantified using the nonparametric bootstrap. The performance of our method is assessed with simulated data. We then describe how our method could be used for species tree inference in larger taxon samples, and demonstrate its utility using datasets for Sistrurus rattlesnakes and for soybeans. The method to infer the phylogenetic relationship among quartets is implemented in the software SVDquartets, available at www.stat.osu.edu/∼lkubatko/software/SVDquartets.