Detecting Rare Genetic Disorders at Population Scale

Kyle Retterer, MS, Chief Data Science Officer at Geisinger, will discuss scalable methods for genomic-first ascertainment of rare disorders and the results of applying these methods to a healthcare population of over 200,000 study participants.

Overview of Presentation

• Genomic-first ascertainment was employed to evaluate 218,680 participants for 2,701 high-confidence rare genetic disorders (RGD)
• We identified 2.5% of participants with high-confidence RGD molecular findings plus 0.7% with possible molecular findings from compound-heterozygous or novel loss-of-function variants
• We developed an ensemble method for assessing diagnostic fit (DxFit) and found that only 15.0% - 21.1% of high-confidence molecular positives had evidence of a DxFit in their medical record
• The low rate of clinical correspondence suggests that genomic ascertainment is more sensitive than clinical methods and that the penetrance of RGDs may be significantly overestimated when studied only in clinically-ascertained populations