The Data Insights team at Roche Sequencing Solutions attended the Curating the Clinical Genome meeting in Hinxton, UK this year. The conference was a great opportunity to hear from individuals actively involved in use of genetics in the clinic and leading initiatives to define and improve the state-of-the-art for the field.
In the keynote presentation of this month’s Bio-IT World Conference and Expo, Dr. Heidi Rehm, Chief Laboratory Director at Partners Healthcare Personalized Medicine, emphasized the need to publicly share genetic variant information to advance patient care. She candidly told the story where her lab at Partners Healthcare reported the result of a genetic test for a fetus at risk for Noonan syndrome as “pathogenic”, only to find out later that there was conflicting evidence and the gene mutation should not be ruled as pathogenic in certain ethnic groups. When she contacted the physician to present the new finding, she was told that the parents had terminated the pregnancy.
The Bina AAiM software is designed to provide deeper insight into variants identified by next-generation sequencing. After uploading a vcf file to the software, variants are annotated against a number of databases that provide information such as predicted pathogenicity, disease association, population frequency, and more. While at small scale, the operation is usually a simple lookup in a text file or a database, performing this annotation for large datasets and at scale is a challenge. Further, due to the variability in the way variants are represented across data sources and VCF files, finding all matching variants requires careful standardization of how variants are represented. Variant normalization and lift-over to a consistent reference genome are two standardization steps that the Bina AAiM software applies in order to find the maximum number of matching variants, which we will explore in this blog post and the next.