We found that by 6 post-conception weeks, most cardinal classes' fate decisions have occurred (remarkably early!).

The exception? The dI4/dI5 lineage decision, which shows similar dynamics between chick and human embryos, demonstrating deep evolutionary conservation.

We extended our approach to human embryos, developing a method to barcode cultured spinal cord slices (CS16-17, ~6 weeks post-conception). We traced lineages at single-cell resolution, opening a window into previously inaccessible human development.

We validated this dual-route mechanism in human ESC-derived cultures. Clonal labeling at different timepoints (day 14 vs day 20) revealed stage-dependent lineage output, matching the chick in vivo patterns!

We zoomed into dI4 & dI5, key (and very abundant!) players in pain and itch processing. They arise through DUAL developmental routes:
• Fast-differentiating unifated dp4 prog. → dI4 only
• Slower bifated dp4/5 prog. → both dI4 & dI5

Lineage & differentiation timing can shape subtype diversity

Oligodendrocyte progenitors (OPCs) share clones with motor neurons and p3 progenitors, supporting the temporal competence switch model where pMN progenitors first make neurons, then switch to making glia

Individual progenitors generate neurons across multiple temporal waves while staying within their lineage subdivision. Spatial identity persists even as temporal competence changes - revealing how space and time interact during development

The nested, hierarchical structure suggests a patterning mechanism: instead of linear monotonic morphogen thresholds, progenitor domains arise via sequential binary decisions that progressively restrict cell fate potential

This made us speculate that lineage architecture co-evolved with the CNS. Lineage structure resembles the ancestral bilaterian “columnar” plan with dorsal/intermediate/ventral stripes. dI6–V1 subdivision (locomotor coordination) may be a vertebrate-specific innovation enabling refined motor control.

These lineage subdivisions align with circuit function. The major lineage split separates dI5/dI6, corresponding to the classic alar-basal division separating sensory-processing & motor-control systems. Lineage not only shapes cell identity but might also anticipate aspects of circuit organisation

When we analysed barcode sharing, we had the first surprise: instead of just 11 independent progenitor domains, we found FIVE broad lineage subdivisions that sit hierarchically above the canonical domains. These subdivisions constrain what cell types progenitors can generate.

We adapted LARRY (Lineage And RNA RecoverY) barcoding for chick embryos & human spinal cord explants. We tracked ~1,500 clones simultaneously capturing single-cell transcriptomes - linking cell identity to lineage.
[Barcode sharing between mature cell types = same origin/one labelled progenitor]

The Question: We know the spinal cord has 11 molecularly distinct progenitor domains generating distinct neurons, but how are these linked by actual LINEAGE relationships? When do progenitors commit to specific cardinal neuron fates, and how early is that commitment resolved?

This was an incredible opportunity to meet old friends, make new ones and listen to diverse and beautiful science...and -why not- go back to pretty Bad Nauheim. Thanks to Didier for creating such a friendly and collaborative community!