Unidad de Bioestadística
The Biostatistics Unit provides an expert approach to statistical problems that arise in biomedical research. The aim of this unit is to promote quality experimental design and research planning, and ensure the use of the most appropriate methods and statistical techniques for data analysis. Thus, resulting in projects with an optimal approach capable of drawing the right conclusions and end successfully.
The functions of the Biostatistics Unit are to provide support and statistical advice to researchers and to collaborate in the development and statistical design of experiments, as well as in the processing, analysis and interpretation of experimental data. It also provides training in the field of biostatistics.
The main tool used for data analysis at the Biostatistics unit is R, which is an open source programming language and software environment for statistical computing and graphics. R is highlighted by its flexibility and adaptability to any kind of research project as well as the compatibility with other standard programs (SPSS, Stata, SAS, Excel, Access, SQL...), offering a wide variety of potent tools for data analysis.
- Workstation with quad-core processor at 3.4 Ghz and 16 GB RAM (64-bit Windows).
- Laptop workstations with dual-core processors at 2.2 Ghz and 2 GB RAM (64-bit Windows).
- Statistical analysis and bioinformatics software: R, Bioconductor, Python and G*Power.
The multidisciplinary combination of professionals with knowledge in disciplines such as biology, statistics, clinics, etc. provides a unique global perspective for dealing with the most challenging experimental design or data analysis.
SERVICES AND EXPERTISSE
The Biostatistics Unit has its services at internal and external customer disposal. It provides a personalized service, expanding experimental design options and offering different methodological approaches to enhance research possibilities.
The Biostatistics Unit offers a wide variety of customized services to enhance basic or applied research:
• Study design (research projects and clinical trials)
• Review of experimental procedures
• Sample size determination and statistical power
• Classical statistical analysis
• Development of explanatory and predictive statistical models
• Analysis of omics data
• Advanced analysis of complex data
• Results interpretation
• Design and preparation of graphs and tables
• Review and editing of articles
• R programming (algorithms, automation analysis, etc.)
Details of the statistical techniques offered:
Study design: Sample size estimation, generation of randomization lists, methodological advice, statistical power optimization.
Classical statistics: Descriptive statistics, parametric tests, non-parametric tests, multiple comparisons, exploratory analysis.
Statistical modeling: Predictive models, inference models, logistic regression, poisson regression, gamma regression, multivariate analysis.
Survival analysis: Kaplan-Meier, Cox regression, parametric models, competing risk models, frailty models.
Analysis of omics data: FDR calculation and q-values, partial least squares (PLS and PLS-DA), LASSO, Elastic Net.
Machine Learning: Neural networks, support vector machines, random forest, boosting and bagging.
Unsupervised analysis: Clustering techniques, Principal component analysis (PCA), self-organizing maps (SOM).
Complex data models: Mixed models for non-independent measures, robust models, non-linear models (GAM), censored regression, Bayesian statistics.
Simulation and resampling methods: Monte-Carlo, bootstrapping, jacknife, cross-validation, permutation tests.