#25 · Tools & infrastructure

Linking neuronal metabolism, callosal genetics, and imaging pipelines to brain disease

The common thread is computational integration of heterogeneous brain data, from imaging to genome-scale models, to relate structural and molecular variation to disease phenotypes. TRAECR standardizes PET preprocessing for statistical modeling through registration, MRI co-registration, and COMBAT/RAVEL harmonization. A TBI preprint pursues phenotype discovery across heterogeneous multi-site segmentations, motivated by the limits of Glasgow Coma Scale outcomes. On the molecular side, iNeuron-GEM, a curated genome-scale metabolic network for human neurons, is coupled with snRNA-seq from ROSMAP and SEA-AD to resolve neuronal subtype-specific metabolic changes in neurodegenerative and neuropsychiatric disease. A GWAS meta-analysis of 46,685 European individuals, using AI-extracted corpus callosum morphometry, maps the genetic architecture of callosal subregions and their neuropsychiatric associations. Notable is the pairing of harmonized imaging pipelines with cell-type-resolved metabolic and genetic models.