Related Experiment Video
Updated: Oct 11, 2025
09:19
Spatial Profiling of Protein and RNA Expression in Tissue: An Approach to Fine-Tune Virtual Microdissection
Published on: July 6, 2022
5.1K
Statistical testing in transcriptomic-neuroimaging studies: A how-to and evaluation of methods assessing spatial and
Yongbin Wei1, Siemon C de Lange1,2, Rory Pijnenburg1
1Department of Complex Trait Genetics, Center for Neurogenomics and Cognitive Research, Amsterdam Neuroscience, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands.
Human Brain Mapping
|December 4, 2021
Summary
Integrating gene transcriptomic and neuroimaging data requires robust statistical models. This study highlights the critical need for gene specificity analysis to avoid reporting nonspecific effects in brain research.
Area of Science:
- Neuroscience
- Genomics
- Biostatistics
Background:
- Multiscale integration of gene transcriptomic and neuroimaging data is crucial for understanding brain organization in health and disease.
- Accurate statistical evaluation is essential to differentiate true associations from random effects.
Purpose of the Study:
- To emphasize the necessity of statistical models that assess "gene specificity" in transcriptomic-neuroimaging analyses.
- To illustrate the challenges and potentials of these integrated analyses.
- To address the issue of nonspecific effects in current methodologies.
Main Methods:
- Discussion of statistical models for transcriptomic-neuroimaging data.
- Illustrative examples of common analyses.
- Simulations to quantify nonspecific effects.
Main Results:
- Nonspecific effects can account for up to 60% of reported associations.
- Ordinary linear regression shows spatial and gene specificity in less than 5% of cases.
- Gene specificity is crucial for avoiding erroneous conclusions.
Conclusions:
- Robust statistical models incorporating gene specificity are vital for reliable transcriptomic-neuroimaging research.
- Current methods may overestimate associations due to a lack of specificity.
- A toolbox for null models is provided to aid researchers.
