Computational Statistics workshops

Translate biological hypotheses into causal questions using potential outcomes and directed graphs. The practical programme identifies confounding, mediation and collider bias and implements matching, propensity scores, inverse weighting, g-computation and doubly robust estimation. It then develops causal machine-learning approaches, including forests, targeted learning, uncertainty quantification and conformal prediction, before applying mediation analysis to multi-omics integration and reproducible R/Python workflows. Sessions: 2–6 November, 14:00–18:00 daily, Europe/Berlin (CET, UTC+01:00), with Slack discussion between sessions. Live online training; fee €480. For advanced MSc, doctoral and postdoctoral researchers with R/Bioconductor, probability, statistics and regression experience.

This course develops a metabolomics investigation from experimental design through data preparation and statistical interpretation. Lectures, practical work and group discussion connect analytical chemistry and data collection with reproducible analysis. Exercises cover R data handling and visualisation, xcms processing of untargeted mass-spectrometry data, quality control, missing values and compound annotation. The final module compares univariate tests, multiple-testing correction, principal-component methods, partial least squares and random forests for metabolomic matrices. Sessions: 16, 17, 18 and 20 November, 09:30–13:30, Europe/Berlin (CET, UTC+01:00); no session on 19 November. Live online training; fee €480. Familiarity with R is assumed; prior metabolomics experience is useful.

Develop Bayesian analyses of biological data from prior specification and posterior estimation to computational diagnostics and multimodal integration. Practical work covers conjugate and non-conjugate models, Gibbs and Hamiltonian Monte Carlo, convergence checks, hierarchical regression, regularisation and variational inference. Later sessions compare data-fusion strategies and Bayesian factor models while quantifying uncertainty in integrated biomedical datasets. Sessions: 30 November–4 December, 14:00–18:00 daily, Europe/Berlin (CET, UTC+01:00), with Slack discussion between meetings. Live online training; fee €480. For advanced MSc, doctoral and postdoctoral researchers with R/Bioconductor, probability, statistics and regression experience.

Estimate species occurrence while accounting for imperfect detection. Participants prepare repeated survey histories, distinguish occupancy from observation processes and fit single-season models with ecological and survey covariates. Exercises compare two-stage estimation in tsoccu with likelihood-based analysis in unmarked, then develop predictions, uncertainty estimates, diagnostics and reproducible reporting. The final day introduces dynamic, multispecies, spatial and false-positive extensions through ecological applications. Sessions: 1–4 December, 09:00–13:00 daily, Europe/Berlin (CET, UTC+01:00). Live online training; fee €420. No prior occupancy-modelling experience is required; practical work is in R.

Learn regression models that estimate nonlinear relationships from data using splines. Starting from generalised linear models, the course develops fitting and visualisation with mgcv and gratia, explains how smoothing controls overfitting, and covers model checking, selection and inference. Further exercises introduce interacting smooths, hierarchical models with group and individual effects, and posterior simulation, using ecological and environmental examples. Sessions: 7–10 December, 14:00–20:00 daily, Europe/Berlin (CET, UTC+01:00). Live online training; fee €480. R/RStudio fluency is required. Familiarity with generalised linear models and fixed/random effects is useful.

Solve shape-analysis problems in R using measurements, landmarks and outlines. The programme develops data organisation and image processing before geometric transformations, alternative superimposition methods and Procrustes analysis. Participants visualise shape differences, test allometry and multivariate hypotheses, and compare principal-component, discriminant and regression approaches. Outline analysis introduces several Fourier representations and their statistical interpretation; participants can bring their own morphological datasets. Live sessions: 10, 11, 14, 15, 16, 17 and 18 December, 12:00–17:00, Europe/Madrid (CET, UTC+01:00). Total 40 hours: 35 live plus five independent. English-language live online course; Fee €495 until 31 October, then €565; Ambassador rates €396/€452. VAT included. Capacity: 16. Basic R, multivariate statistics and geometric morphometrics are recommended; this is not an introductory morphometrics course. Registration requires acceptance; payment follows confirmation. Recordings are provided, but live attendance is required for a certificate.

Learn to build and evaluate machine-learning workflows in R with tidymodels and tidyverse. The workshop moves from problem definition, dataset splitting and preprocessing to decision trees, random forests and gradient boosting. Practical work covers cross-validation, hyperparameter tuning, regression and clustering, with model-building demonstrations, live coding, quizzes and individual or group exercises. Participants should already be comfortable with R, basic data analysis and statistics; machine-learning experience is not required. Sessions on 8, 15, 22 and 29 January 2027, 16:00–19:30 Madrid time (CET). Fourteen live hours plus eight hours of independent exercises. Fee including VAT: EUR 546 through 31 October 2026, then EUR 610; Ambassador Institution rates EUR 436.80 and EUR 488. Limited to 15 participants. Acceptance is confirmed before payment. Delivered live online in English with practical exercises. Live attendance is required for a certificate.

This course introduces Markov and latent-state models for longitudinal observations. Participants learn model specification, estimation and interpretation, including covariate extensions and responses that are continuous, categorical or incomplete. Worked examples and exercises use the LMest package in R and connect the statistical assumptions with applied research questions. It is designed for academic and industry participants with little prior probability training; basic R familiarity is helpful. The four online sessions run 14:00–17:00 Berlin local time. The listed package fee is €420.

Build and evaluate short-term ecological forecasts by repeatedly linking observations, models and predictions. A biodiversity challenge uses shrub, rabbit and dung-beetle data from Doñana with remote-sensing measurements. Participants process time series and satellite indices, compare autoregressive, hierarchical, correlational and mechanistic models, propagate uncertainty and assess forecasts against benchmarks. Reproducible R workflows lead to a submitted forecast and discussion of how predictions support conservation and adaptive management. Live sessions: 25–29 January, 14:00–18:30 daily, Europe/Madrid (CET, UTC+01:00). Total 27.5 hours: 22.5 live plus five independent. English-language live online course; Fee €236 until 30 November, then €290; Ambassador rates €188.80/€232. VAT included. Capacity: 25. Basic R and an ecology or environmental-science background are required; forecasting experience is unnecessary. Registration requires acceptance; payment follows confirmation. Recordings are provided, but live attendance is required for a certificate.

Researchers learn the assumptions and implementation principles behind machine-learning models used in cancer genomics. The five-day programme develops hidden Markov approaches for copy-number inference, mixture models for tumour subclones, hierarchical models for mutation signals and autoencoders for single-cell data. Lectures alternate with discussion, coding practice and a final project, emphasizing inference of latent biological states from noisy or incomplete measurements. Participants may discuss suitable datasets they own. Online teaching runs 14:00–19:00, Monday through Friday.

Combine complementary omics measurements to identify biological pathways and model cellular behaviour. Lectures and practicals compare supervised feature selection and integration using LASSO, PLS, LDA, mixOmics and DIABLO with unsupervised MOFA models and deep-learning autoencoders. The final day covers dimensionality reduction, graph-based integration, batch correction and multimodal single-cell analysis using Seurat and weighted nearest neighbours. Participants learn to select and implement an integration strategy suited to their research question. Live online, 15–17 February, 14:00–19:00 Europe/Berlin (CET). Fee €450. Basic Unix awareness and beginner-level R or Python are expected.

The course teaches causal reasoning for ecological and evolutionary studies where controlled experiments may be impractical. Participants translate biological hypotheses into directed graphs, identify confounding and selection structures, and fit Bayesian structural equation models in R. Practical work covers posterior diagnostics, mediation, mixed response types, hidden causes, competing models and phylogenetic dependence. The programme emphasizes what observational evidence can and cannot establish, with opportunities to develop analyses for participants’ own systems. Online sessions run 14:00–18:00 Berlin time. The course fee is €480.

Use population genomic data to investigate structure, gene flow, demographic history and natural selection. The workshop starts with short-read quality control, alignment and variant calling, then examines supervised and unsupervised population-structure inference and introgression. Further sessions compare coalescent, composite-likelihood and approximate Bayesian demographic models, genome-wide association and selection scans, and genotype–environment associations. Landscape-genomic exercises test isolation by distance, environmental effects and barriers to gene flow. Live online, 22–26 February, 14:00–19:00 Europe/Berlin (CET). Fee €530. Intended for graduate students, postdocs and researchers with genetics and evolutionary biology knowledge; programming experience, Unix and R are expected.

Develop phylogenetic comparative methods from their probabilistic foundations using Julia. Participants implement likelihood and Bayesian inference, including Markov chain Monte Carlo, and study Brownian motion, continuous-time Markov chains and birth–death processes. These tools are applied to trait evolution, biogeography and diversification, including the effects of environmental and biotic interactions. The practical course is intended for researchers familiar with programming, probability and phylogenetics; prior Julia experience is not required. Sessions on 1, 3, 5, 8, 10 and 12 March 2027, 14:00–18:00 Madrid time (CET). Fee including VAT: EUR 492 through 31 December 2026, then EUR 576; Ambassador Institution rates EUR 393.60 and EUR 460.80. Limited to 18 participants. Acceptance is confirmed before payment. Delivered live online in English with practical exercises. Live attendance is required for a certificate.

Foundation models and synthetic data are increasingly used in scientific applications, yet the reliability of their outputs is often uncertain. This workshop investigates how model predictions and generated data can be incorporated into statistical inference without losing trustworthy uncertainty estimates. Researchers from several disciplines will study the implications of black-box models for inference and decision-making and develop methods for assessing and improving reliability.