Partitioning of Individual Autozygosity into Multiple Homozygous-by-Descent Classes

Functions to identify Homozygous-by-Descent (HBD) segments associated with runs of homozygosity (ROH) and to estimate individual autozygosity (or inbreeding coefficient). HBD segments and autozygosity are assigned to multiple HBD classes with a model-based approach relying on a mixture of exponential distributions. The rate of the exponential distribution is distinct for each HBD class and defines the expected length of the HBD segments. These HBD classes are therefore related to the age of the segments (longer segments and smaller rates for recent autozygosity / recent common ancestor). The functions allow to estimate the parameters of the model (rates of the exponential distributions, mixing proportions), to estimate global and local autozygosity probabilities and to identify HBD segments with the Viterbi decoding. The method is fully described in Druet and Gautier (2017) .

Bayesian Multivariate Analysis of Summary Statistics

Multivariate tool for analyzing genome-wide association study results in the form of univariate summary statistics. The goal of 'bmass' is to comprehensively test all possible multivariate models given the phenotypes and datasets provided. Multivariate models are determined by assigning each phenotype to being either Unassociated (U), Directly associated (D) or Indirectly associated (I) with the genetic variant of interest. Test results for each model are presented in the form of Bayes factors, thereby allowing direct comparisons between models. The underlying framework implemented here is based on the modeling developed in "A Unified Framework for Association Analysis with Multiple Related Phenotypes", M. Stephens (2013) .

Functions to Simplify the Use of 'glmnet' for Machine Learning

Provides several functions to simplify using the 'glmnet' package: converting data frames into matrices ready for 'glmnet'; b) imputing missing variables multiple times; c) fitting and applying prediction models straightforwardly; d) assigning observations to folds in a balanced way; e) cross-validate the models; f) selecting the most representative model across imputations and folds; and g) getting the relevance of the model regressors; as described in several publications: Solanes et al. (2022) , Palau et al. (2023) , Sobregrau et al. (2024) .

Programmatic Utilities for Manipulating Formulas, Expressions, Calls, Assignments and Other R Objects

These utilities facilitate the programmatic manipulations of formulas, expressions, calls, assignments and other R language objects. These objects all share the same structure: a left-hand side, operator and right-hand side. This packages provides methods for accessing and modifying this structures as well as extracting and replacing names and symbols from these objects.

Extended Model Formulas

Infrastructure for extended formulas with multiple parts on the right-hand side and/or multiple responses on the left-hand side (see ).

Phase I/II CRM Based Drug Combination Design

Implements the adaptive designs for integrated phase I/II trials of drug combinations via continual reassessment method (CRM) to evaluate toxicity and efficacy simultaneously for each enrolled patient cohort based on Bayesian inference. It supports patients assignment guidance in a single trial using current enrolled data, as well as conducting extensive simulation studies to evaluate operating characteristics before the trial starts. It includes various link functions such as empiric, one-parameter logistic, two-parameter logistic, and hyperbolic tangent, as well as considering multiple prior distributions of the parameters like normal distribution, gamma distribution and exponential distribution to accommodate diverse clinical scenarios. Method using Bayesian framework with empiric link function is described in: Wages and Conaway (2014) .

Spatial Generalised Linear Mixed Models for Areal Unit Data

Implements a class of univariate and multivariate spatial generalised linear mixed models for areal unit data, with inference in a Bayesian setting using Markov chain Monte Carlo (MCMC) simulation using a single or multiple Markov chains. The response variable can be binomial, Gaussian, multinomial, Poisson or zero-inflated Poisson (ZIP), and spatial autocorrelation is modelled by a set of random effects that are assigned a conditional autoregressive (CAR) prior distribution. A number of different models are available for univariate spatial data, including models with no random effects as well as random effects modelled by different types of CAR prior, including the BYM model (Besag et al., 1991, ) and Leroux model (Leroux et al., 2000, ). Additionally, a multivariate CAR (MCAR) model for multivariate spatial data is available, as is a two-level hierarchical model for modelling data relating to individuals within areas. Full details are given in the vignette accompanying this package. The initial creation of this package was supported by the Economic and Social Research Council (ESRC) grant RES-000-22-4256, and on-going development has been supported by the Engineering and Physical Science Research Council (EPSRC) grant EP/J017442/1, ESRC grant ES/K006460/1, Innovate UK / Natural Environment Research Council (NERC) grant NE/N007352/1 and the TB Alliance.

Mixed GAM Computation Vehicle with Automatic Smoothness Estimation

Generalized additive (mixed) models, some of their extensions and other generalized ridge regression with multiple smoothing parameter estimation by (Restricted) Marginal Likelihood, Generalized Cross Validation and similar, or using iterated nested Laplace approximation for fully Bayesian inference. See Wood (2017) for an overview. Includes a gam() function, a wide variety of smoothers, 'JAGS' support and distributions beyond the exponential family.

Simultaneous Inference in General Parametric Models

Simultaneous tests and confidence intervals for general linear hypotheses in parametric models, including linear, generalized linear, linear mixed effects, and survival models. The package includes demos reproducing analyzes presented in the book "Multiple Comparisons Using R" (Bretz, Hothorn, Westfall, 2010, CRC Press).

One-to-One Feature Matching

Statistical methods to match feature vectors between multiple datasets in a one-to-one fashion. Given a fixed number of classes/distributions, for each unit, exactly one vector of each class is observed without label. The goal is to label the feature vectors using each label exactly once so to produce the best match across datasets, e.g. by minimizing the variability within classes. Statistical solutions based on empirical loss functions and probabilistic modeling are provided. The 'Gurobi' software and its 'R' interface package are required for one of the package functions (match.2x()) and can be obtained at < https://www.gurobi.com/> (free academic license). For more details, refer to Degras (2022) "Scalable feature matching for large data collections" and Bandelt, Maas, and Spieksma (2004) "Local search heuristics for multi-index assignment problems with decomposable costs".