CoLab: Accurate and Scalable WGS Variant Prioritization in the NICU and for Phenotype-Free Newborn Screening

Thursday, October 16, 3:15pm - 3:45pm (EDT)

Time shown in-04:00 America, New York

Dr. Mark Yandell, a professor at the University of Utah, showcases the capabilities of Fabric GEM AI and its development across various applications. It is demonstrating its value in contexts ranging from the Neonatal Intensive Care Unit (NICU) to Newborn Screening (NBS), as well as in population datasets and real-world initiatives like GeneKids.

In NICU applications where phenotype data is available, GEM achieves a mode and median rank of 1, placing the true positive genotype among the top five candidates in about 95% of cases. Typically, fewer than three cases require human review.

NBS presents even greater challenges due to the lack of phenotype data and the necessity for extremely low false-positive rates. Despite these conditions, GEM maintains high accuracy by effectively prioritizing single nucleotide variants (SNVs), insertions and deletions (indels), and copy number variants (CNVs), regardless of whether the causative variants are listed in ClinVar.

Additionally, we will present new data from the GeneKids program, where reanalysis using Fabric GEM is revealing diagnoses in previously negative cases, underscoring the potential of these tools to cost-effectively enhance diagnostic yields.

Add to Calendar 2025/10/16 15:15:00 2025/10/16 15:45:00 America/New_York CoLab: Accurate and Scalable WGS Variant Prioritization in the NICU and for Phenotype-Free Newborn Screening Dr. Mark Yandell, a professor at the University of Utah, showcases the capabilities of Fabric GEM AI and its development across various applications. It is demonstrating its value in contexts ranging from the Neonatal Intensive Care Unit (NICU) to Newborn Screening (NBS), as well as in population datasets and real-world initiatives like GeneKids.

In NICU applications where phenotype data is available, GEM achieves a mode and median rank of 1, placing the true positive genotype among the top five candidates in about 95% of cases. Typically, fewer than three cases require human review.

NBS presents even greater challenges due to the lack of phenotype data and the necessity for extremely low false-positive rates. Despite these conditions, GEM maintains high accuracy by effectively prioritizing single nucleotide variants (SNVs), insertions and deletions (indels), and copy number variants (CNVs), regardless of whether the causative variants are listed in ClinVar.

Additionally, we will present new data from the GeneKids program, where reanalysis using Fabric GEM is revealing diagnoses in previously negative cases, underscoring the potential of these tools to cost-effectively enhance diagnostic yields. ASHG - Theater #2 false MM/DD/YYYY 30 OPAQUE aTlxRIWVvzXukcEPdmbf176894

ASHG - Theater #2

Estephania MR, ereese@fabricgenomics.com

Link: