LIRICAL

LIRICAL applies a likelihood-ratio framework to integrate Human Phenotype Ontology-encoded phenotypes with genomic variant data (VCF from diagnostic gene panels, exome sequencing, or whole-genome sequencing) to prioritize candidate Mendelian disease diagnoses and estimate posttest probabilities.

Key Features:

• Likelihood Ratio Framework: Employs a likelihood ratio paradigm to quantify how observed phenotypes and genotypes support each candidate diagnosis.
• Posttest Probability Estimation: Computes posttest probabilities for candidate diagnoses by combining phenotypic and genotypic likelihood ratios.
• Phenotype Contribution Analysis: Evaluates the contribution of each observed Human Phenotype Ontology (HPO) term to diagnosis prioritization.
• Predicted Pathogenicity Assessment: Assesses the predicted pathogenicity of observed genotypes to inform genotype likelihoods.
• HPO and VCF Integration: Integrates phenotypic data encoded with HPO and genetic data from diagnostic gene panels, exome sequencing, or whole-genome sequencing provided in VCF format.
• Robustness to Data Noise: Accounts for phenomic and genomic noise and demonstrates resilience to typical forms of data variability via simulations.

Scientific Applications:

• Rare Mendelian disease diagnosis: Prioritizes candidate diagnoses in genomic diagnostics for rare Mendelian diseases, placing the correct diagnosis within the top three ranks in 92.9% of cases with a mean posttest probability of 67.3%.
• Interpretable diagnostic metrics: Provides interpretable metrics beyond ranked gene lists to support diagnostic interpretation using combined phenotypic and genotypic evidence.

Methodology:

Integrates HPO-encoded phenotypic data with genomic variants from sequencing results (VCF), applies likelihood ratios to compute phenotype- and genotype-based likelihoods, calculates posttest probabilities, evaluates per-HPO contribution to rankings, assesses predicted pathogenicity of variants, and assesses robustness to phenomic and genomic noise via simulations.