Skip to main content

Conference Program

Subpage Hero

Loading

Using Genomic Databases to Estimate Rare Disease Prevalence: Application to Disorders of Propionyl-CoA Oxidation 

Biochemical/Metabolic and Therapeutics
  • Primary Categories:
    • Metabolic Genetics
  • Secondary Categories:
    • Metabolic Genetics
Introduction:
Inborn errors of propionate metabolism are a group of autosomal recessive, genetically heterogenous, severe and potentially lethal disorders detected via elevated propionylcarnitine on newborn screening (NBS). As part of public and private partnerships, including the PaVe-GT and Bespoke Gene Therapy Consortium efforts, the development of AAV-gene therapy is underway for isolated methylmalonic acidemia caused by variants in MMUT or MMAB and propionic acidemia due to deficiency of propionyl-CoA carboxylase, encoded by PCCA or PCCB. Data from NBS studies suggest these conditions should be classified as rare diseases however the exact prevalence of each disorder is unknown and complicated by genetic heterogeneity and the lack of a nationwide NBS registry.

Methods:
We first interrogated exome/genome sequencing datasets, gnomAD v2.1.1 and SVs 2.1 for variants in MMUT, MMAB, PCCA and PCCB and identified pathogenic or likely pathogenic (P/LP) ClinVar variants. Variants that were conflicting, of uncertain significance, or not annotated were cross-referenced against HGMD Professional version 2022.3 and reviewed by a board-certified molecular geneticist to score pathogenicity according to the ACMG guidelines. To augment classification of variants, we included additional disease-causing variants present in our cohort of participants evaluated at the NIH Clinical Center (NCT00078078, NCT02890342). Gene carrier rate (GCR) was determined by the sum of the minor allele frequency (MAF) of all P/LP variants. Carrier frequency and predicted incidence were calculated using the Hardy-Weinberg equation.

Results:
For MMUT, 153 P/LP variants were identified with GCR of 0.0024, equating to an estimated carrier frequency of 1/209 and predicted prevalence of 1/175,849. Variants were individually rare and missense changes accounted for about half. We identified 37 P/LP variants in MMAB, yielding a GCR of 0.0007. We calculated an estimated carrier frequency of 1/677 and prevalence of 1/1.8 million. The c.556G>T (p.Arg186Trp), was the most common variant, consistent with observations that it accounts for ~30-40% of alleles in MMAB cohorts. For propionic acidemia, we identified 93 variants in PCCA and 85 variants in PCCB with similar GCRs of 0.0013 and calculated a combined prevalence of ~1/319,000. Most variants in PCCB were missense whereas 70% of variants in PCCA were loss of function. One commonly reported insertion/deletion variant in PCCB c.1218_1228delinsTAGAGCACAGGA may be underestimated given poor alignment and incorrect annotation.

Conclusion:
To our knowledge, there is no official regulatory guidance on the use of genomic databases, which are massively expanding across the globe, to assist with population estimates of rare disease prevalence and incidence, which are required to report to the FDA during the development of rare disease therapeutics. Although genome-based analysis of the prevalence of MMUT, MMAB, PCCA and PCCB pathogenic variants represents a conservative estimate of disease allele frequency, the subsequent predicted birth prevalence calculated using the Hardy-Weinberg equation closely agrees with the number of infants diagnosed with MMA and PA in a cohort of ~4.5 million babies who underwent NBS in California (PMID: 32778825). For example, we estimate a prevalence of MMUT-type MMA as~1/175,849 while NBS data reports ~1/110,000. Our analyses demonstrates that there is concurrence between epidemiologic and genomic data and should be useful to support FDA regulatory designations. Extending this approach to other rare genetic disorders that lack data from newborn screening but have promising therapies under development is feasible.

Agenda

Sponsors