Systematic Evaluation of Competing Brain Transcriptomic Representations Reveals Reciprocal Patterns Across...

Area of Science:

• Neuroscience
• Genomics
• Bioinformatics

Background:

• Brain states are often modeled as a molecular continuum.
• This assumption lacks robust transcriptomic validation against alternative models.

Purpose of the Study:

• Compare a global-axis model with a low-dimensional reciprocal model for brain transcriptomic organization.
• Evaluate model performance across diverse contexts like exercise, alcohol, stress, aging, and neurodegeneration.

Main Methods:

• Benchmarked models using signed-effect correlations, permutation analyses, and held-out reconstruction.
• Utilized cross-study brain transcriptomic data with directional contrast labels.
• Performed region-specific human bulk, cellular, single-nucleus, spatial, and chromatin analyses.

Main Results:

• The global-axis model showed weak correlations and did not outperform permutation null expectations.
• A rank-1 latent profile within the reciprocal model better reconstructed held-out genes than a module 1/module 2 (M1/M2) partition.
• Human analyses revealed asymmetric patterns and poor cross-stratum reproducibility, emphasizing context and cohort dependence.

Conclusions:

• A low-dimensional reciprocal representation offers a more effective exploratory framework for brain transcriptomic organization.
• Findings underscore the critical importance of context dependence, cohort dependence, and biological heterogeneity.