Learning outcomes

At the end of the course unit, students will be able to select, apply and interpret statistical methods appropriate for the analysis of biological data. They will be able to translate a biological question into a statistical problem, select an appropriate method, verify its assumptions, perform the analysis using R, interpret the results and present them appropriately.

The course develops these skills through two complementary modules focusing on linear statistical modelling and multivariate data analysis.

Goals

The course aims to develop both a conceptual and practical understanding of biological data analysis, covering two complementary approaches.

The linear statistical modelling module aims to deepen students’ understanding of linear models and extend it to generalised linear models and linear mixed models. Emphasis is placed on selecting a model appropriate to the biological question and data structure, understanding its assumptions, and interpreting the results.

The multivariate data analysis module aims to familiarise students with the analysis of datasets containing several variables simultaneously. It introduces different approaches for visualising and summarising the structure of multivariate data, investigating similarities and differences among observations or groups, grouping observations, and assigning observations to groups.

In both modules, emphasis is placed on statistical reasoning and practical application using R, rather than solely on formal knowledge of statistical methods.

Content

The course is organised into two complementary modules.

Module 1 – Linear statistical modelling

  • Review of basic statistical concepts

  • Review of normal linear models

  • Generalised linear models

  • Linear mixed models

Module 2 – Multivariate data analysis

  • Multivariate data and their visualisation

  • Ordination using Principal Component Analysis (PCA)

  • Correspondence Analysis and Multidimensional Scaling

  • Permutation tests on a (dis)similarity matrix

  • Grouping objects: clustering

  • Assigning objects to groups: discriminant analysis

Table of contents

Module 1: Linear statistical modeling


  • Chapter 1: Recap of normal linear models
  • Chapter 2: Generalized linear models (random part, linear predictor, link function; parameter estimation, inference)
  • Chapter 3: Linear mixed models (basics; LMMs as multivariate normal distributions; parameter estimation, inference)

Module 2: Multivariate data analysis


  • Chapter 1: Multivariate data and their visualization
  • Chapter 2: Ordination by Principal Component Analysis (PCA)
  • Chapter 3: Ordination of a contingency table: Correspondence Analysis (CA)
  • Chapter 4: Other ordination techniques: Multidimensional scaling
  • Chapter 5: Grouping objects: clustering
  • Chapter 6: Assigning objects to groups: discriminant analysis


Exercices

  • Practical sessions in a computer lab with R.
  • Solving concrete statistical problems related to ecology.
  • Tutorials to apply theoretical concepts.


Teaching methods

Teaching combines lectures, seminars and computer-based practical sessions. Active student participation is encouraged in both modules.

In the linear statistical modelling module, lectures and exercises progressively develop the use of linear, generalised and mixed models. Seminars also address the application and interpretation of these methods in the context of biological problems.

In the multivariate data analysis module, students learn to explore and visualise multivariate data and apply different methods of ordination, comparison, grouping and classification.

In both modules, exercises develop students’ ability to independently solve a statistical problem: identifying an appropriate analysis, checking its assumptions, performing the analysis using R, interpreting the results and presenting them appropriately.

For the linear statistical modelling module, attendance at practical sessions and seminars designated as mandatory is required. Only duly justified absences are accepted. Any timetable conflict must be reported at the beginning of the academic year, before October. An unjustified absence from a mandatory activity may prevent the student from taking the assessment for this module.

Assessment method

The two modules are assessed separately, with each module contributing 10/20 to the final grade.

Module 1 – Linear Statistical Modelling

Assessment consists of an open-book examination, including exercises on linear mixed models and generalised linear models, carried out using R. Students may consult their syllabus and other authorised materials during the examination. The use of artificial intelligence (AI) tools, including conversational assistants or code-generation tools, is prohibited during the examination.

Module 2 – Multivariate Data Analysis

Assessment consists of an open-book examination that may include multiple-choice questions, open-ended questions, and practical exercises using R on a computer. After the written exam, there will be an oral defence of the answers.

The final grade is the sum of the grades obtained for the two modules. As the final grade must be expressed as a whole number, it is rounded up if both modules are passed (at least 5/10 for each module), and rounded down if this is not the case.

Sources, references and any support material

  • Lecture slides and practical materials available on Moodle / WebCampus.
  • Datasets and R scripts provided.
  • Additional references and resources indicated on the platform.


Language of instruction

English
Training Study programme Block Credits Mandatory
Master in Biology of Organisms and Ecology Finalité approfondie 1 5 Yes
Master in Biology of Organisms and Ecology Finalité didactique 1 5 Yes