Using mixed effects models to analyse longitudinal gene expression can highlight differences between sample groups over time. The most widely used differential gene expression tools are unable to fit linear mixed effect models, and are less optimal for analysing longitudinal data. This package provides negative binomial and Gaussian mixed effects models to fit gene expression and other biological data across repeated samples. This is particularly useful for investigating changes in RNA-Sequencing gene expression between groups of individuals over time, as described in: Rivellese, F., Surace, A. E., Goldmann, K., Sciacca, E., Cubuk, C., Giorli, G., ... Lewis, M. J., & Pitzalis, C. (2022) Nature medicine <doi:10.1038/s41591-022-01789-0>.

Fits high dimensional penalized generalized linear mixed models using the Monte Carlo Expectation Conditional Minimization (MCECM) algorithm. The purpose of the package is to perform variable selection on both the fixed and random effects simultaneously for generalized linear mixed models. The package supports fitting of Binomial, Gaussian, and Poisson data with canonical links, and supports penalization using the MCP, SCAD, or LASSO penalties. The MCECM algorithm is described in Rashid et al. (2020) <doi:10.1080/01621459.2019.1671197>. The techniques used in the minimization portion of the procedure (the M-step) are derived from the procedures of the ncvreg package (Breheny and Huang (2011) <doi:10.1214/10-AOAS388>) and grpreg package (Breheny and Huang (2015) <doi:10.1007/s11222-013-9424-2>), with appropriate modifications to account for the estimation and penalization of the random effects. The ncvreg and grpreg packages also describe the MCP, SCAD, and LASSO penalties.

Several tests for high dimensional generalized linear models have been proposed recently. In this package, we implemented a new test called adaptive sum of powered score (aSPU) for high dimensional generalized linear models, which is often more powerful than the existing methods in a wide scenarios. We also implemented permutation based version of several existing methods for research purpose. We recommend users use the aSPU test for their real testing problem. You can learn more about the tests implemented in the package via the following papers: 1. Pan, W., Kim, J., Zhang, Y., Shen, X. and Wei, P. (2014) <DOI:10.1534/genetics.114.165035> A powerful and adaptive association test for rare variants, Genetics, 197(4). 2. Guo, B., and Chen, S. X. (2016) <DOI:10.1111/rssb.12152>. Tests for high dimensional generalized linear models. Journal of the Royal Statistical Society: Series B. 3. Goeman, J. J., Van Houwelingen, H. C., and Finos, L. (2011) <DOI:10.1093/biomet/asr016>. Testing against a high-dimensional alternative in the generalized linear model: asymptotic type I error control. Biometrika, 98(2).

Cross validation informed Relaxed LASSO, Artificial Neural Network (ANN), gradient boosting machine ('xgboost'), Random Forest ('RandomForestSRC'), Oblique Random Forest ('aorsf'), Recursive Partitioning ('RPART') or step wise regression models are fit. Cross validation leave out samples (leading to nested cross validation) or bootstrap out-of-bag samples are used to evaluate and compare performances between these models with results presented in tabular or graphical means. Calibration plots can also be generated, again based upon (outer nested) cross validation or bootstrap leave out (out of bag) samples. For some datasets, for example when the design matrix is not of full rank, glmnet may have very long run times when fitting the relaxed lasso model, from our experience when fitting Cox models on data with many predictors and many patients, making it difficult to get solutions from either glmnet() or cv.glmnet(). This may be remedied by using the path=TRUE option when calling glmnet() and cv.glmnet(). Within the glmnetr package the approach of path=TRUE is taken by default. When fitting not a relaxed lasso model but an elastic-net model, then the R-packages nestedcv <https://cran.r-project.org/package=nestedcv>, glmnetSE <https://cran.r-project.org/package=glmnetSE> or others may provide greater functionality when performing a nested CV. Use of the glmnetr has many similarities to the glmnet package and it is recommended that the user of glmnetr also become familiar with the glmnet package <https://cran.r-project.org/package=glmnet>, with the "An Introduction to glmnet'" and "The Relaxed Lasso" being especially useful in this regard.

Data sets from the book Generalized Linear Models with Examples in R by Dunn and Smyth.

Penalized methods are useful for fitting over-parameterized models. This package includes functions for restructuring an ordinal response dataset for fitting continuation ratio models for datasets where the number of covariates exceeds the sample size or when there is collinearity among the covariates. The glmnet fitting algorithm is used to fit the continuation ratio model after data restructuring.

Builds a LASSO, Ridge, or Elastic Net model with glmnet or cv.glmnet with bootstrap inference statistics (SE, CI, and p-value) for selected coefficients with no shrinkage applied for them. Model performance can be evaluated on test data and an automated alpha selection is implemented for Elastic Net. Parallelized computation is used to speed up the process. The methods are described in Friedman et al. (2010) <doi:10.18637/jss.v033.i01> and Simon et al. (2011) <doi:10.18637/jss.v039.i05>.

In practical applications, the assumptions underlying generalized linear models frequently face violations, including incorrect specifications of the outcome variable's distribution or omitted predictors. These deviations can render the results of standard generalized linear models unreliable. As the sample size increases, what might initially appear as minor issues can escalate to critical concerns. To address these challenges, we adopt a permutation-based inference method tailored for generalized linear models. This approach offers robust estimations that effectively counteract the mentioned problems, and its effectiveness remains consistent regardless of the sample size.

We provide an efficient implementation for two-step multi-source transfer learning algorithms in high-dimensional generalized linear models (GLMs). The elastic-net penalized GLM with three popular families, including linear, logistic and Poisson regression models, can be fitted. To avoid negative transfer, a transferable source detection algorithm is proposed. We also provides visualization for the transferable source detection results. The details of methods can be found in "Tian, Y., & Feng, Y. (2023). Transfer learning under high-dimensional generalized linear models. Journal of the American Statistical Association, 118(544), 2684-2697.".

This package provides diagnostic graphic tools for GLMs, beta-binomial regression model (estimated by VGAM package), beta regression model (estimated by betareg package) and negative binomial regression model (estimated by MASS package). Since most of functions implemented in glmxdiag already exist in other packages, the aim is to provide the user unique functions that work on almost all regression models previously specified. Details about some of the implemented functions can be found in Brown (1992) <doi:10.2307/2347617>, Dunn and Smyth (1996) <doi:10.2307/1390802>, O'Hara Hines and Carter (1993) <doi:10.2307/2347405>, Wang (1985) <doi:10.2307/1269708>.

Fit linear models to overdispersed count data. The package can estimate the overdispersion and fit repeated models for matrix input. It is designed to handle large input datasets as they typically occur in single cell RNA-seq experiments.

This package provides a function for fitting a penalized constrained continuation ratio model using the glmpath algorithm and methods for extracting coefficient estimates, predicted class, class probabilities, and plots as described by Archer and Williams (2012) <doi:10.1002/sim.4484>.

This package provides a variable selection approach for generalized linear mixed models by L1-penalized estimation is provided, see Groll and Tutz (2014) <doi:10.1007/s11222-012-9359-z>. See also Groll and Tutz (2017) <doi:10.1007/s10985-016-9359-y> for discrete survival models including heterogeneity.

Recursive partitioning based on (generalized) linear mixed models (GLMMs) combining lmer()/glmer() from lme4 and lmtree()/glmtree() from partykit'. The fitting algorithm is described in more detail in Fokkema, Smits, Zeileis, Hothorn & Kelderman (2018; <DOI:10.3758/s13428-017-0971-x>). For detecting and modeling subgroups in growth curves with GLMM trees see Fokkema & Zeileis (2024; <DOI:10.3758/s13428-024-02389-1>).

Specification, analysis, simulation, and fitting of generalised linear mixed models. Includes Markov Chain Monte Carlo Maximum likelihood and Laplace approximation model fitting for a range of models, non-linear fixed effect specifications, a wide range of flexible covariance functions that can be combined arbitrarily, robust and bias-corrected standard error estimation, power calculation, data simulation, and more. See <https://samuel-watson.github.io/glmmr-web/> for a detailed manual.

Fit Generalized Linear Models to continuous and count outcomes, as well as estimate the prevalence of misrepresentation of an important binary predictor. Misrepresentation typically arises when there is an incentive for the binary factor to be misclassified in one direction (e.g., in insurance settings where policy holders may purposely deny a risk status in order to lower the insurance premium). This is accomplished by treating a subset of the response variable as resulting from a mixture distribution. Model parameters are estimated via the Expectation Maximization algorithm and standard errors of the estimates are obtained from closed forms of the Observed Fisher Information. For an introduction to the models and the misrepresentation framework, see Xia et. al., (2023) <https://variancejournal.org/article/73151-maximum-likelihood-approaches-to-misrepresentation-models-in-glm-ratemaking-model-comparisons>.

This package implements Bayesian spatial and spatiotemporal models that optionally allow for extreme spatial deviations through time. glmmfields uses a predictive process approach with random fields implemented through a multivariate-t distribution instead of the usual multivariate normal. Sampling is conducted with Stan'. References: Anderson and Ward (2019) <doi:10.1002/ecy.2403>.

Optimal design analysis algorithms for any study design that can be represented or modelled as a generalised linear mixed model including cluster randomised trials, cohort studies, spatial and temporal epidemiological studies, and split-plot designs. See <https://github.com/samuel-watson/glmmrBase/blob/master/README.md> for a detailed manual on model specification. A detailed discussion of the methods in this package can be found in Watson, Hemming, and Girling (2023) <doi:10.1177/09622802231202379>.

Set of tools for the statistical analysis of data using: (1) normal linear models; (2) generalized linear models; (3) negative binomial regression models as alternative to the Poisson regression models under the presence of overdispersion; (4) beta-binomial and random-clumped binomial regression models as alternative to the binomial regression models under the presence of overdispersion; (5) Zero-inflated and zero-altered regression models to deal with zero-excess in count data; (6) generalized nonlinear models; (7) generalized estimating equations for cluster correlated data.

This package provides a Bayesian model selection approach for generalized linear mixed models. Currently, GLMMselect can be used for Poisson GLMM and Bernoulli GLMM. GLMMselect can select fixed effects and random effects simultaneously. Covariance structures for the random effects are a product of a unknown scalar and a known semi-positive definite matrix. GLMMselect can be widely used in areas such as longitudinal studies, genome-wide association studies, and spatial statistics. GLMMselect is based on Xu, Ferreira, Porter, and Franck (202X), Bayesian Model Selection Method for Generalized Linear Mixed Models, Biometrics, under review.

This package provides two functions that generate source code implementing the predict function of fitted glm objects. In this version, code can be generated for either C or Java'. The idea is to provide a tool for the easy and fast deployment of glm predictive models into production. The source code generated by this package implements two function/methods. One of such functions implements the equivalent to predict(type="response"), while the second implements predict(type="link"). Source code is written to disk as a .c or .java file in the specified path. In the case of c, an .h file is also generated.

This package provides a collection of tools to streamline the process of fitting elastic net models with glmnet. In addition to providing a formula interface, it also features a function cva.glmnet to do crossvalidation for both α and λ, as well as some utility functions.

Allows users to fit a cosinor model using the glmmTMB framework. This extends on existing cosinor modeling packages, including cosinor and circacompare', by including a wide range of available link functions and the capability to fit mixed models. The cosinor model is described by Cornelissen (2014) <doi:10.1186/1742-4682-11-16>.

This package provides functions to calculate predicted values and the difference between the two cases with confidence interval for lm() [linear model], glm() [generalized linear model], glm.nb() [negative binomial model], polr() [ordinal logistic model], vglm() [generalized ordinal logistic model], multinom() [multinomial model], tobit() [tobit model], svyglm() [survey-weighted generalised linear models] and lmer() [linear multilevel models] using Monte Carlo simulations or bootstrap. Reference: Bennet A. Zelner (2009) <doi:10.1002/smj.783>.