Research integrity is not just about extreme misconduct — it is also about the human biases that insidiously push science from responsible practice toward questionable habits, undermining reproducibility and credibility.

This seminar offers concrete tools and strategies applicable at the bench and largely relevant to any researcher, whatever their field—including if you attended last year, as content has been updated.

In this year’s keynote lecture, Joe Garner (Stanford University) proposes a powerful mindset shift in experimental design illustrated by examples from animal research, but broadly applicable beyond the animal lab: moving from “what did we control?” to “what did we choose to ignore — and at what cost?”.

When, where

• Date : Thursday May 28, 1:00–4 :45 PM
• Location : Geopolis, room 1612 (UNIL Mouline)
• Followed by an apéritif

Who should attend?

• Compulsory for CIG PhD students and postdocs who joined since March 21, 2025.
• 0.5 day RESAL credit for continuing education in Animal Experimentation: register
• All welcome —250 participants last year, from across the Lemanic area.

Programme

Introduction : Research integrity – The Good, the Bad and the Ugly — Frédéric Preitner

Ethics at the bench (1): Managing bias in experimental design and data visualization — Frédéric Schutz

Ethics at the bench (2): Beautification, manipulation, and AI: ethics of scientific images — Romain-Daniel Gosselin

The FBM Ombudsman introduces himself —Christian Kern

Keynote lecture : Changing the scientific mindset in biomedical studies — Joe Garner (Stanford University)

About the keynote speaker

Joe Garner is Professor of Comparative Biomedicine at Stanford University, where his work on animal models and experimental design has made him one of the leading voices on reproducibility in biomedical research. His keynote challenges researchers to move beyond “what did we control?” and ask instead: “what did we choose to ignore — and at what cost?”

Further information about Joe Garner’s work

Beyond3Rs: https://med.stanford.edu/beyond3rs.html
Stanford website: med.stanford.edu/profiles/Joseph_Garner
Papers: scholar.google.com/citations?user=JThS8LMAAAAJ
Mouse ethogram: www.mousebehavior.org

Talk abstracts

1. Keynote lecture: Changing the scientific mindset in biomedical studies (Joe Garner). This seminar explores a necessary shift in perspective in biomedical research: moving from the question “What have we controlled for?” to a more critical one: “What have we chosen to ignore, and at what cost?” Using examples from studies employing animal models, it highlights six common blind spots that can compromise scientific conclusions: ignored differences in the biology of the animal model, ignored characteristics of the modeled human disease, ignored limitations of the measurements used, ignored environmental factors, ignored impacts of animal welfare, and ignored principles of experimental design and statistics. The seminar demonstrates how recognizing these implicit assumptions can improve the validity, reproducibility, and ethical foundations of biomedical research.
2. Ethics at the bench (1): Experimental design and data presentation (Frédéric Schutz). This seminar examines sources of bias in research, with examples from mouse studies. It shows how bias can arise even when researchers act in good faith, and why safeguards such as controls, randomization, and blinding are essential. It also emphasizes that the goal of science is not to prove hypotheses but to challenge them rigorously. The seminar further explores how uncontrolled environmental variability—such as cage effects, where animals sharing a cage are not statistically independent—can influence experimental outcomes. Finally, it addresses bias in data visualization and reporting, highlighting how misleading graphs or ambiguous error bars can distort interpretation and emphasizing the need for transparent presentation of data.
3. Ethics at the bench (2): Beautification, manipulation, and AI: ethics of scientific images (Romain Gosselin). This seminar examines ethical issues related to the handling and manipulation of scientific images in biomedical research, mostly focusing on immunoblotting. It emphasizes that digital images are full-fledged data that must be handled with the same rigor as any other dataset. The presentation discusses how choices made during image acquisition, processing, analysis, and reporting can introduce bias. It also reviews common questionable practices, such as excessive contrast adjustment, saturation, cropping, or duplication of image elements. Finally, the seminar highlights the concerns raised by the abrupt rise of generative artificial intelligence, which now enable the generation of highly realistic synthetic images indiscernible from genuine images. Implications for research integrity and the need for awareness, good practices, and appropriate detection tools are discussed.
4. Presentation of the FBM Ombudsman (Christian Kern). The Ombudsman of the Faculty of Biology and Medicine (FBM) at UNIL will introduce his role and mission. He will explain the mechanisms for reporting suspected scientific misconduct, which may arise in all types of research—clinical or basic—and may involve human subjects, animals, or biological material. In cases of suspected misconduct, the Ombudsman acts as an advisory and guidance authority. The Ombudsman will present several examples of situations that may constitute breaches of scientific integrity in clinical practice or laboratory research.