Thread by @thatdnaguy

#ASHG21 ML: Mischa Lundberg.
Integrative analysis of human genetic association studies, single cell transcriptomics, and knockout mouse models identifies cell types and genes involved in skeletal disease.

#ASHG21 ML: Lack of 'omics data for skeletal disorders. Use murine 'omics to try and fill this gap since the data aren't always available for human samples.

#ASHG21 ML: Calls sets of genes expressed in cell types a profile, and that profiles are conserved between species in many cases.

#ASHG21 ML: Identify profiles of cells from mouse femurs (bone cells) that overlap with monogenic skeletal disease causing genes [i.e. try to flag potential bone effector cells from their profiles].

#ASHG21 ML: Cells from 3 compartments in mouse femurs. 133,942 single cells. Cluster cells based on similarity of profiles. 34 distinct clusters. Each cluster defined by distinct DE genes.

#ASHG21 ML: ~420 known monogenic disease causing genes. Hypergeometric test for enrichment relative to human/mouse orthologs. 3 robustly enriched cell types. Subgroup analysis show enrichment in high bone mass and bone growth disorders.

#ASHG21 ML: Looked at enrichment for polygenetic skeletal traits, such as height. MAGMA gene-based tests for association with each of ~20k genes. Mapped to mouse orthologs. Gene set analysis - is the set of genes that define a cell type more related to the trait genes than chance

#ASHG21 ML: Looked for mouse knockouts that have bone mineral density changes. Genes associated with these changes in mice are enriched in genes that define cell types in their profiles.

#ASHG21 ML: Genes that have an effect on BMD are often genes with restricted expression and help define the cell type they are expressed in.

#ASHG21 ML: SLC9A3R2 is an example that has restricted expression and is a potential effector.

#ASHG21 ML: SLC9A3R2 knockout mice have altered bone marrow density. So the method can potentially identify novel effectors that can be confirmed in mouse, cell line, and other models.