Decoding Disease: Can Genetics Unlock New Interventions? | Future of Health Summit 2024

Current State of Genetic Discovery

• Knowledge of disease-causing genes has expanded from approximately 8–10 genes 25 years ago to 10,000–12,000 identified variants today.
• Genetic discoveries are becoming routine, but translating these findings into diagnostics, devices, and treatments remains a primary bottleneck.
• Experts suggest the field is emerging from a "hype cycle" bottom, analogous to the trajectories of monoclonal antibodies, AI, and the microbiome.
• A critical path to treating common diseases involves deconstructing them into rare, monogenic subsets, a strategy validated by the histological subtyping of breast cancer (currently 80 subtypes).
• Cancer treatment faces a "narrow window of intervention" for aggressive, metastatic forms (specifically prostate cancer) where distinguishing between indolent and metastatic disease is difficult.

Bottlenecks in Translating Genetics to Care

• Insurance and Access:
  • Medicare does not currently recognize genetic counselors as independent providers, forcing patients to pay out-of-pocket.
  • Medicaid coverage for genetic counseling varies significantly across states.
  • A new CPT code exists, but hospitals still lack reimbursement mechanisms to hire genetic counselors.
  • Legislation, specifically the "Access to Genetic Counselor Services Act," is required to ensure widespread insurance coverage.
• Provider Literacy:
  • Many healthcare providers lack deep understanding of gene therapy and genomics, necessitating the embedding of genetic counselors into multidisciplinary teams (e.g., cardiology, neurology).
  • The volume of ongoing clinical trials (approx. 20,000 for cancer) makes it impossible for individual physicians to stay informed without AI or crowdsourced decision support.
• Data Representation:
  • Current genomic data is heavily skewed toward individuals of European ancestry, limiting the predictive accuracy of polygenic risk scores for African, Hispanic, and Latino populations.
  • Reference databases like TCGA are insufficient for determining risk alleles in African ancestry men, specifically regarding aggressive prostate cancer.

Technological Advancements and Future Directions

• AI and Computational Tools:
  • Machine learning is increasingly used to analyze vast genomic data to direct targeted therapy and identify tumor heterogeneity.
  • There is caution regarding Large Language Models (LLMs) potentially hitting a utility plateau or "eating their own tail" due to training on model-generated data.
  • AI is expected to have a near-term impact on analyzing coding variations and predicting protein structural changes to identify disease-causing missense mutations.
• Model Systems:
  • Drug discovery is shifting toward reducing complex diseases into rare, high-impact mutations that can be studied in human cell-based models (iPSCs), organoids, and in silico simulations.
  • In silico models are expected to complement animal models (mice, zebrafish) and cell lines to better recapitulate human-specific biological processes, particularly in neuroscience.
  • The cost of sequencing is dropping to approximately $200, enabling statewide population screening programs (e.g., Alabama's Catalyst program, Minnesota's Health Partners).

Specific Disease Area Insights

• Psychiatry and Mental Health:
  • Mental health is the leading cause of global disability, yet genetic progress lags behind cancer due to the difficulty of studying the human brain.
  • Strong-acting rare variants may offer a "shortcut" to understanding complex disorders like bipolar disorder and schizophrenia by fracturing them into coherent biological subsets.
  • Genetic findings challenge current diagnostic categories; for instance, ACAP11 loss-of-function mutations linked to bipolar disorder also appear in schizophrenia, suggesting potential repurposing of lithium therapy.
  • Genetic counseling in this space focuses on gene-environment interactions to prevent patient powerlessness (if viewed as purely genetic) or guilt (if viewed as purely environmental).
• Cancer:
  • Clinical practice in oncology is currently ahead of the basic science curve regarding sequencing adoption.
  • Epigenetic changes (e.g., DNA methylation in blood) are becoming viable non-invasive biomarkers for monitoring disease and therapeutic response.
  • A "tumor board" crowdsourcing model in rural Maine has demonstrated improved survival by connecting local diagnoses to global clinical trials.

Patient Navigation and Engagement

• Behavioral Impact: Genetic testing can drive behavioral modification (diet, exercise) more effectively than standard medical advice due to personal relevance.
• Diagnostic Challenges: Patients with rare diseases often act as "genetic navigators" themselves, utilizing social media to find other individuals with matching mutations to attract pharmaceutical research attention.
• Recommendations for Patients:
  • Patients should not rely on direct-to-consumer testing alone; results must be contextualized with family history and clinical data.
  • The primary step for the public is to consult a physician or genetic counselor to interpret genetic information within a broader health context.
  • Participation in large-scale research initiatives (e.g., NIH All of Us program) is encouraged to diversify data sets and accelerate discovery.

Timeline and Expectations

• Systematic analysis of genetic variants across all human diseases using biobanks and electronic health records is projected to be achievable within a 10–15 year horizon.
• Population-level health outcomes tracking for genomic screening programs (e.g., prenatal, adult) is expected to begin yielding data within the next 5 years.
• The goal of the Stanley Center is to sequence 250,000 individuals with bipolar disorder within the next 5–7 years to achieve necessary statistical power.