Genetics

This post covers four recent papers that use single-cell and single-nucleus data to ask which cells hold out in MS, none of them using the autoimmune animal model, and finds the answer indexed every time. It covers the single-cell eQTL Mendelian randomization framework across 14 immune cell types, where APOM and MSH5 came out protective for MS while AHI1 ran in opposite directions depending on the lineage its expression was measured in, and reports that paper's own sensitivity analyses. It then covers the integrative single-nucleus and spatial study of human MS white matter in which an IKZF1-linked transcription factor module shifts in microglia and shifts most at the lesion rim, which the authors read as the inflammatory hotspot where remyelination is hindered. It covers the Piezo2 study in which single-nucleus data placed a mechanosensitive channel in late oligodendrocytes, its deletion disrupted myelin compaction specifically in small-calibre axons, and PIEZO2 was reduced in optic nerve lesions from MS patients. It closes on the methodological floor under all of it: how transcripts get assigned to cells in postmortem human brain, and the fact that microglia, neurons and neurovascular cells are the types most sensitive to that choice.