Disease-causing variant recommendation system for clinical genome interpretation with adjusted scores for artefactual variants

Abstract

In the process of finding the causative variant of rare diseases (RD), accurate assessment and prioritization of genetic variants is essential. Although quality control (QC) of genetic variants is strictly performed, the presence of artefactual variants in the remaining set of variants can deteriorate the process. Variant QC and prioritization have been treated as separate processes, leading to limited efficiency and risk of misdiagnosis. We developed a disease-causing variant recommendation system that integrates quality control into variant prioritization by adjusting scores for artefactual variants. We confirmed that the QC-related features of the variants contribute to a significant performance improvement. For genomic data from 2,878 patients with rare disorders, the recall rate of finding causative variants was 0.961 for the top 5 ranked variants. We also found that our system recognized the anomaly of QC-related features, so that the scores of artifactual variants to be disease-causing were assessed relatively low. Integration of variant QC and prioritization help reduce the risk of misdiagnosis based on artefactual variants and increase the effectiveness of clinical genome interpretation.

Exome and genome sequencing for diagnosing patients with suspected rare genetic disease

“…Then, the variants are prioritized by how similar the patient's clinical symptoms are to the reported symptoms of the disease associated with the gene the variant occurred in. Since this process is labor-intensive and time-consuming, there are now variant recommendation algorithms being developed using artificial intelligence (AI) technology [38][39][40]. Top-k, the frequency the diagnostic variant is found in the top-k variants, is a commonly used measure to assess the performance of the algorithms.…”