Analytical Performance of a Genome-Phenome Analyzer for Use in a Clinical Laboratory (P5.133)

Objectives: To evaluate the analytical performance of a genome-phenome analyzer software program that uses Bayesian pattern matching and rank orders candidate genes based on their variant pathogenicity scores and pertinence to clinical findings. Background: Neurogenetic diseases affect 6[percnt] to 10[percnt] of the population and are the most common indication for exome sequencing in children. The burden of integrating genome and phenome data is placed on a limited number of individuals with experience in genomic medicine. Decision support software may broaden the capability to interpret such data. Methods: Exome sequencing was performed on de-identified genomic DNA representing different neurogenetic disorders (n=19) and polycystic kidney disease (n=1) with known pathogenic variants identified by Sanger sequencing. Clinical information for the 20 cases was entered independently into the genome-phenome analyzer by 4 board-certified genetic counselors and 4 geneticists (Total = 160 entries). The software input consisted of patient summaries and exome .vcf files that included fields for standard prediction tools and publicly available datasets. The software output was used to compare raters and calculate performance characteristics. Results: The genome-phenome analyzer had a sensitivity of 93[percnt] and a positive predictive value (PPV) of 83[percnt] (n=160). The majority were true positive results (78[percnt]; n=125) with low false positive (16[percnt]; n=26) and false negative (6[percnt]; n=9) rates. Concordance between raters ranged from 75[percnt] to 85[percnt], with 3 cases missed by all raters (15[percnt]). A correct result was predicted in 80[percnt] of cases when 2 or more raters agreed. Conclusions: The evaluated genome-phenome analyzer has high sensitivity for ranking genes pertinent to the diagnosis and may aid clinical laboratories in the interpretation of exome sequencing results. Clinical judgment remains important in determining the significance of the software output, especially in patients with a negative genetic test result who do not have a neurogenetic disease. Disclosure: Dr. Higgins has received personal compensation for activities with Quest Diagnostics. Dr. Wang has received personal compensation for activities with Quest Diagnostics as an employee. Dr. Jaremko has received personal compensation for activities with Quest Diagnostics as an employee. Dr. Batish has received personal compensation for activities with Quest Diagnostics as an employee. Dr. York has received personal compensation for activities with Quest Diagnostics as an employee.