Figure 1. Integrating Mendelian pedigrees with single cell epigenomic data.

a. Schematic depicting subset of human cMNs and their targeted muscles. cMN3 (blue) = oculomotor nucleus which innervates the inferior rectus, medial rectus, superior rectus, inferior oblique, and levator palpebrae superior muscles; cMN4 (purple) = trochlear nucleus which innervates the superior oblique muscle; cMN6 (green) = abducens nucleus which innervates the lateral rectus muscle (bisected); cMN7 (pink) = facial nucleus which innervate muscles of facial expression; cMN12 (black) = hypoglossal nucleus which innervates tongue muscles. Corresponding CCDDs for each cMN are listed under diagram and color coded. CFEOM: congenital fibrosis of the extraocular muscles; CP: congenital ptosis; FNP: fourth nerve palsy; DRS: Duane retraction syndrome; MBS: Moebius syndrome; CFP: congenital facial palsy.

b. Overview of the experimental and computational approach. i) Generating cell type-specific chromatin accessibility profiles. Brightfield and fluorescent images of e10.5 Isl1^MN:GFP embryo (left) from which cMNs are microdissected (yellow dotted lines, dissociated, FACS-purified (middle), followed by scATAC and data processing (right; red and blue lines represent adapters, black line represents DNA, orange cylinders represent nucleosomes, grey pentagons represent Tn5). ii) WGS of 270 CCDD pedigrees (left; 899 individuals; sporadic and inherited cases) followed by joint variant calling, QC, and Mendelian variant filtering (right). iii) Integrating genome-wide non-coding variant calls with epigenomic annotations for variant nomination (top). To aid in variant interpretation, we identify cognate genes (2^nd row), aggregate candidate variants, generate functional variant effect predictions (3^rd row), and validate top predictions in vivo (bottom).

c. UMAP embedding of single cell chromatin accessibility profiles from 86,089 GFP-positive cMNs, sMNs, and their surrounding GFP-negative neuronal tissue colored based on GFP reporter status (left, GFP-positive green, GFP-negative grey), sample (middle, with sample key under UMAP) and cluster (right, with cluster annotations in Supplementary Table 3). Gridlines in middle UMAP apply to left and right UMAPs as well. The inset shows the relative proximity of Cluster 2 cells dissected from the same cell type (cMN7 e10.5) from different technical and biological replicates.

d. Heatmap depicting the proportions of dissected cells within each of the 23 major clusters. Homogeneity/completeness metrics are shown for GFP-positive versus GFP-negative clusters. cMN6 and cMN7 are in close spatial proximity and are commonly co-dissected.