Purchase this article with an account.
Laura Bryant, Olga Lozynska, Jeannette Bennicelli, Tomas S Aleman, Albert M Maguire, Jean Bennett; Semi masked analysis reduces bias in whole exome sequencing analysis. Invest. Ophthalmol. Vis. Sci. 2017;58(8):2753.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Whole Exome Sequencing (WES) is a powerful tool that provides detailed information about the genetic mutations in a patient population, but the amount of information generated can make the analysis time consuming and difficult. We used a semi-masked directed approach to efficiently identify pathogenic mutations in patients with retinal degeneration.
In an IRB-approved study, we collected genomic DNA from 85 subjects with various forms of retinal degeneration who either self-referred or who were diagnosed at UPenn. WES was performed by the Penn Genome Frontiers Institute. We used a python script to generate a list of candidate genes for each subject consisting of any retinal degeneration gene with a low frequency variant in the subject, defined as a variant with either an unknown frequency or an allele frequency of less than 0.05. After the list of candidate genes was made, we prioritized the genes based on the family history and patient diagnosis. Segregation analysis of the mutation with the family pedigree was performed when possible.
We identified the pathogenic mutations in 56 of the subjects. The mutations were in 29 different genes. Some of the subjects had a dominant mutation when the family history suggested a recessive form of inheritance. Two subjects had syndromic forms of retinal degeneration when they had previously been diagnosed with only retinal degeneration.
Generating a list of candidate genes without taking into account the specific diagnosis or family history allowed us to easily identify pathogenic mutations in 56 of the 85 subjects we sequenced. Screening for mutations in the genes with known importance to retinal function, we were able to quickly eliminate more than half of the subjects from further analysis. Retinal phenotype is often used as an initial filtering step to help make the dataset a more manageable size. Contrarily, we found that limiting the analysis based on the diagnosis filtered out too much information and had the potential to eliminate disease causing variants from the analysis. It also relies on the initial diagnosis being completely accurate, which can be difficult when the disease phenotypes have significant overlap. Using only retinal degeneration as the initial phenotype resulted in a manageable list of potentially pathogenic variants while limiting the amount of bias we introduced with our filters.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.
This PDF is available to Subscribers Only