Rare Variant Gene-Based Case-Control Burden Testing in Hypogonadotropic Hypogonadism Using Publically Available Control Data

Presentation Number: SUN 469
Date of Presentation: April 2nd, 2017

Michael Guo1, Lacey C Plummer2, Joel N Hirschhorn1 and Margaret Flynn Lippincott*2
1Boston Children's Hospital, Boston, MA, 2Massachusetts General Hospital, Boston, MA


Background: Isolated GnRH Deficiency (IGD) is characterized as failure to enter or progress through puberty and occurs in between 1:30,000-1:125,000 births. Over 35 mendelian loci have been identified; however, together these ‘known IGD genes’ account for only 1/3 of the cases in our large cohort ~1,800 patients. The inheritance patterns for the known IGD genes are mixed: autosomal dominant (AD), autosomal recessive (AR), and X-linked. The majority of IGD cases appear to be sporadic given the reproductive consequences of this disorder. Given the severity of this disease we hypothesized that IGD probands would be enriched for rare deleterious variants in novel genes affecting the biology of GnRH neurons compared to publically available controls.

Aims: To develop gene-based burden testing models of whole exome sequencing (WES) data using publically available control data and IGD cases based on an autosomal dominant inheritance pattern. To validate these models using known IGD genes.

Methods: Whole-exome sequencing was done on 395 IGD cases. Ancestry was assigned by using principal components analysis. This cohort was enriched for IGD-causing variants in PROKR2 (a positive control). The publically available Exome Aggregation Consortium (ExAC) WES data was used as a control cohort. For the model, each data set was restricted to rare (MAF <0.1%), putatively deleterious (nonsense, frameshift, essential splice, or missense that are computationally predicted to be damaging). These multiple variants were then collapsed by gene and the gene-based burden of variants was compared between cases and controls 2x2 contingency table tests.

Results: The model revealed known IGD genes: FGFR1 (12 variants in cases p<7.4E-9); PROKR2, the positive control, (8 variants in cases, p<2.3E-5) and TACR3 (7 cases, p<2.6E-5). FGFR1, PROKR2, and TACR3 were in the top 5 genes in the model, although they did not all meet exome-wide significance (p<2.5E-6). Other genes that meet exome wide significance, but whose role in IGD is unknown, are being investigated as potential new IGD genes in model systems.

Conclusions: We report the ability to use publically available WES data to design rare variant case-control gene-based burden testing in a rare disease model. This rare variant burden testing was validated by its ability to recapitulate genes known to cause IGD: PROKR2 (a positive control), FGFR1, and TACR3. As a result, novel genes passing the burden testing are being evaluated for their possible role in IGD.


Disclosure: JNH: Principal Investigator, Pfizer, Inc.. Nothing to Disclose: MG, LCP, MFL