The mutational constraint spectrum quantified from variation in 141,456 humans

Area of Science:

• Human genetics
• Genomics
• Population genetics

Background:

• Genetic variants offer insights into gene function and essentiality.
• Loss-of-function variants are informative but challenging to analyze due to errors and low frequencies.

Purpose of the Study:

• To aggregate large-scale human sequencing data for variant analysis.
• To identify and analyze high-confidence predicted loss-of-function variants.
• To classify human genes based on their tolerance to inactivation.

Main Methods:

• Aggregated 125,748 exomes and 15,708 genomes into the Genome Aggregation Database (gnomAD).
• Filtered variants to identify 443,769 high-confidence predicted loss-of-function variants.
• Developed an improved model of human mutation rates for gene classification.

Main Results:

• Identified a substantial number of high-confidence loss-of-function variants.
• Classified human protein-coding genes based on their tolerance to inactivation.
• Validated the gene classification model using data from model organisms and engineered human cells.

Conclusions:

• The gnomAD database provides a valuable resource for studying gene function.
• The gene inactivation tolerance spectrum improves gene discovery for diseases.
• This approach enhances the power to identify genes associated with both common and rare human diseases.