Kisang Kwon, a Research Scientist at 3billion's Bioinformatics team will present methods to improve the accuracy of variant classification when diagnosing rare disease patients from underrepresented populations at ASHG 2022.
One of the most important pieces of information used to categorize variant classification is Allele Frequency (AF) using large public databases such as gnomAD (Genome Aggregation Database). These databases still lack racial/ethnic diversity for variant classification, particularly for Asian/African populations which are underrepresented in the existing databases.
Consequently, the variant pathogenicity classification could be more accurate when additional information on these underrepresented populations is collected. Based on the extracted AF information from the actual 20,455 WES patient data analyzed by 3billion, it shows that the genetic variant with high frequency found from a specific Asian population was not found in gnomAD in several cases.
When researchers discover a novel variant using only gnomAD, there could be a risk of overestimating it as pathogenic since it appears as a very rare variant. On the other hand, when AF information of the population to which the patient belongs is used appropriately, it can be seen as a high-frequency (common) variant and be classified as benign. There are real-life cases of Asian/African patients, where the variants are classified as pathogenic when they are not.
Kwon stated, “In order to improve the diagnosis of patients with rare diseases in population groups that are underrepresented in public databases such as gnomAD, it is important to gather data for these groups. 3billion has data from a diverse population that is underrepresented in gnomeAD, with Asian and African patient populations accounting for 70%, which improves the diagnosis of rare diseases for these populations.”