Hypothesis-free phenotype prediction within a genetics-first framework.

Area of Science:

• Genomics and Bioinformatics
• Human Genetics
• Computational Biology

Background:

• Cohort-wide sequencing studies identify rare variants as the largest category, even in coding regions.
• Current associative methods offer limited understanding of how rare genetic variants influence complex phenotypes.
• Established tools often fail to identify genetic causes for certain developmental disorders.

Purpose of the Study:

• To develop and present an ab initio, genetics-first method for molecular knowledge-based interpretation of coding variants.
• To analyze exome-wide non-synonymous variants for organism and cellular level phenotypes.
• To identify plausible genetic causes for developmental disorders and other phenotypes missed by standard methods.

Main Methods:

• A knowledge-based approach utilizing protein domains and ontologies (function and phenotype).
• Consideration of all coding variants, irrespective of allele frequency.
• Reverse approach for variant interpretation applied to a direct-to-consumer genotype cohort.

Main Results:

• Identification of plausible genetic causes for developmental disorders that eluded other methods.
• Generation of molecular hypotheses for the causal genetics of 40 distinct phenotypes.
• Demonstration of additional discoveries possible beyond standard genetic analysis tools.

Conclusions:

• A knowledge-based, genetics-first approach can significantly enhance variant interpretation beyond allele frequency.
• This method provides a powerful tool for uncovering genetic underpinnings of complex phenotypes, including rare diseases.
• The system enables further discovery from existing genetic data, complementing standard analytical pipelines.