DeepBD: A Grounded Agentic Workflow for Variant Prioritization and Diagnosis of Genetic Birth Defects

2026-06-23

AI Summary

This study addresses the challenge of prioritizing and diagnosing pathogenic genetic variants in hereditary birth defects, particularly when fetal or infant phenotypic information is incomplete and evidence sources are heterogeneous. The authors propose a decoupled “reasoning-with-evidence” agent workflow that separates and synergistically coordinates evidence integration, tool invocation, and large language model (LLM)-assisted diagnosis. Their approach combines a pretrained evidence engine, rule-based knowledge, sequence- and variant-effect representations, phenotype-conditioned biological context, and specialized tool modules to enable complementary fusion and precise ranking of multi-source, heterogeneous evidence. Evaluated on an internal cohort of 18,622 cases, the method achieves Recall@1/3/5/10 of 0.658/0.882/0.912/0.929, significantly outperforming baselines including Exomiser, DeepRare, and LLM-based reranking approaches.

Abstract

Birth defects are a major cause of fetal loss, neonatal morbidity and long-term disability. In the subset with suspected genetic etiologies, exome and genome sequencing have moved many cases from variant detection to post-sequencing interpretation: clinicians must rank patient-specific candidate variants under incomplete fetal or infant phenotypes and heterogeneous evidence from population genetics, variant-effect prediction, gene-disease validity, phenotype ontologies, cellular and pathway context, protein structure and clinical literature. We present DeepBD, a grounded agentic workflow for variant prioritization and diagnostic interpretation of genetic birth defects. DeepBD organizes the workflow into LLM-assisted case structuring, a pretrained evidence engine, specialist evidence modules and a grounded diagnostic review layer. The evidence engine learns patient-specific variant scores from structured rule evidence, sequence and variant-effect representations and phenotype-conditioned biological context, whereas specialist modules and the agentic layer provide tool-based refinement, candidate-pool review and diagnosis-oriented synthesis from ranked candidates. Developed using an in-house fetal and infant cohort comprising 18,622 cases, DeepBD achieved Recall@1/3/5/10 of 0.658/0.882/0.912/0.929 on an internal held-out solved-case benchmark, outperforming standalone Exomiser, DeepRare and prompted LLM reranking baselines evaluated on Exomiser-derived top-20 candidate variants. Ablation and overlap analyses show that rule evidence, mechanistic context, and specialist refinement provide complementary signals. These findings support a grounded agentic workflow that separates evidence integration, tool-based refinement, and LLM-assisted diagnostic review for retrospective variant prioritization in genetic birth defects.

Problem

• variant prioritization
• genetic birth defects
• diagnostic interpretation
• phenotype-driven analysis
• candidate variant ranking

Innovation

• grounded agentic workflow
• variant prioritization
• evidence integration
• LLM-assisted diagnosis
• genetic birth defects