Make figures / text based around the cell type specificity of obvious facial markers visible from birth

[NathanSkene]

Jai had the figure showing that leaf nodes of HPO are more specific. Bobby to add an extra one showing the leaf nodes are more likely to only have one cell type association.

This principle can be followed through to the idea that facial morphological defects should be cell type specific.

E.g. there is a broad "abnormal lip morphology" which is associated with excitatory, inhibitory neurons, astrocytes and amacrine cells.

But a subphenotype of that, "short philtrum" is specific to excitatory neurons.

And a subphenotype "cleft upper lip" is astrocyte specific.

Can suggest then that obvious facial defects can be used to gain insight into the underlying cell types.

[NathanSkene]

Could then have another figure, investigating whether "short philtrum" is also indicative of particular cognitive deficits.

• What percentage of short philtrum genes are associated with behavioural / cognitive abnormalities?
• Plot the relationship between specificity to neurons (X-axis) & whether it's associated with cognitive abnormality (with bar graphs?)

[NathanSkene]

Point is to show that the more neuronal it is.... the more likely it is to cause cleft lip... and cognitive defects.... and so these are the same genes causing both?

[NathanSkene]

So the genes which are associated with both cleft lip + cognitive should be even more neuronally specific?

[NathanSkene]

Try calculating hypergeometric enrichments (or the bootstrapping method) between cleft upper lip & cognitive phenotypes.... which are shared?

[bschilder]

Demonstrating the ontology level-celltype specificity relationship is done in Fig 2. Though it doesn't show a specific example, I think it's fine since it illustrates the broader point across all phenotypes. https://github.com/neurogenomics/rare_disease_celltyping/blob/master/figures/manuscript_figures_july_2022.pdf