Curating gene alias collisions for the resolution of gene symbol ambiguity
Year: 2024
Type: other
Abstract: Genomic data harmonization is a necessary step in the creation of searchable, linked knowledge for use in genomic research and clinical practice. This is a complicated endeavor due to the diverse representation of gene concepts across resources including health records, genome sequencing data, and clinical trial reports. This challenge is exacerbated by gene nomenclature ambiguity, particularly due to gene alias collisions, where a single gene symbol represents multiple genes. A prominent example of a collision in the NCBI Gene database is the symbol “KRAS”, which appears as the primary gene symbol of the human KRAS gene, and the alias symbol of another gene, NRAS. NRAS and KRAS are distinct and well-studied genes of the RAS gene family that play a significant role in cancer with over a thousand references and more than 70 variants combined in the Clinical Interpretation of Variants in Cancer (CIViC) knowledgebase. The difficulties associated with resolving ambiguity and ensuring accurate understanding of gene symbols limit the effectiveness of data aggregation tools. There are 3,680 (9.1%) records in Ensembl that share an alias with another gene as well as 2,084 (4.8%) records in HGNC and 5,670 (7.5%) records in NCBI Gene. There are 218 (0.5%) records in Ensembl that share a primary gene symbol as alias with another gene as well as 483 (1.1%) records in HGNC and 1,712 (2.3%) records in NCBI Gene. To improve gene data harmonization, we are working to develop a resource for resolving ambiguous gene symbols found in the biomedical literature and genomic knowledgebases. We curate the relationships between the gene symbols and gene concepts, providing structured semantic data to support automated symbol disambiguation. Our work lays the foundation for evidence-driven protocols for resolving automated gene symbol matching and to improve gene harmonization workflows in knowledge aggregation resources such as DGIdb (dgidb.org) and MetaKB (search.cancervariants.org).
Institution: The Ohio State University
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: gold
OpenAlex ID: W4416504617