Figure Legends
Fig 1: Variant filtering and annotation workflow. The analysis
pipeline included filtering variants based on sample and variant-level
criteria for quality control. The pipeline also included annotating
variants with gene names, control population frequencies, and CADD phred
scores so that variant counts could be collapsed over a gene’s coding
region and so that the analysis could focus on likely deleterious, rare
variants. Abbreviations: VCF = variant call format, GATK = Genome
Analysis Toolkit, dbNSFP = Database of Non-synonymous Single-nucleotide
Variant Functional Predictions version 4.0b1a, gnomAD = Genome
Aggregation Database version 2.
Fig 2: Distribution of Affected Genes per Subject. This figure
compares Mexican-American (MA) myelomeningocele subjects to European
ancestry (EA) myelomeningocele subjects by illustrating how many
disrupted genes the each subject possessed. Green circles represent MA
myelomeningocele data and blue squares represent EA myelomeningocele
data. (A) The horizontal axis measures the number of genes that are
disrupted by qualifying variants. The vertical axis depicts the number
of subjects who possess that number of disrupted genes. (B) The
horizontal axis counts only the number of genes that indicated nominally
significant risk in the mutational burden analysis. The vertical axis
shows the number of subjects with that number of disrupted genes. The
vertical dotted lines from either graph represent the median number of
disrupted genes possessed by subjects, with green lines representing MA
and blue representing EA medians. Both populations followed a similar
distribution for how many genes are disrupted per subject, complete with
a similar median. However, the Mexican Americans tended to have more
nominally significant disrupted genes.
Fig 3: Altered WNT trafficking. The proteins PORCN,
SDC1, and CPE are all involved in WNT ligand trafficking. Deleterious
variants in PORCN may prevent WNT’s acetylation which is necessary for
WNT to leave the endoplasmic reticulum. Loss of PORCN function would
stop WNT from leaving the cell. CPE prevents WNT from binding cell
surface receptors on the target cell. Loss of CPE function may
indirectly increase WNT’s effect on the target cell. Proteoglycans like
SDC1 have been implicated as regulators in WNT distribution, though the
exact mechanism is not yet known.
Fig 4: Altered Planar Cell Polarity. DVL2 ,PRICKLE2 , and FUZ genes were each disrupted in the
myelomeningocele populations. The FZD-DVL complex is necessary to
establish planar cell polarity. DVL proteins are also needed to
translocate FUZ to cilia. FUZ is essential for the development of a
cilium. PRICKLE inhibits the formation of the FZD-DVL complex.
Fig 5: Altered β-catenin. A visual summary of the WNT/β-catenin
cascade including genes disrupted in the myelomeningocele populations.
The level of β-catenin, which is coded for by the human homologCTNNB1 , is regulated by many proteins in the WNT ligand’s target
cell. The proteins CSNK1G, DVL2, DDB1, PSMD3, PTPRU, Fermitin 2 (coded
by FERMT2 ), CPE, PPP2R1A (not shown), and SOSTDC1 contribute to
β-catenin’s regulation and all have higher mutational burdens in one of
the myelomeningocele populations compared to gnomAD controls.