8

Reflections on the Workshop

“This has been an unusual meeting,” said Kent Lloyd in the final session of the workshop. While the workshop presentations demonstrated how far we have come in precision medicine and modeling, there is still more to be done. Lloyd identified four persistent challenges for precision medicine research:

• The growth in translational research has increased the incidence of exposing animal-based research that lacks reproducibility, which is not a good predictor of human response or disease.
• There have been only modest gains in successful treatments; slow and expensive drug pipelines often end in failure in diverse patient populations.
• The path from genotype to mechanism to phenotype is not as simple and straightforward as once thought. Issues such as pleiotropy, heterogeneity, variable penetrance, and sexual dimorphism are complicating—but also enriching—our understanding of how genotypes and phenotypes are related.
• Due to the expansion of genome sequencing, we have identified an exponentially increasing number of variants of unknown significance, while suffering from a lack of sufficient specific biomarkers of disease.

The utility of model organisms for precision medicine, Lloyd said, lies in their predictive value and their proximity to the patient. Referring back to Valle’s presentation he reiterated that three of the best uses of animal

models are to confirm causation, to help us understand pathophysiology, and as surrogates for treatment studies. To those he added a fourth, experimental interrogation.

Based on the presentations and discussions at the workshop, and the needs and obstacles that had been identified, Lloyd developed eight “calls to action.” First, there is a need to complete a comprehensive, encyclopedic functional annotation of the genome. He explained that this project would require collection and integration of data from both humans and animals, and would require experimentation to test hypotheses. While there is much focus on the protein-coding regions of the genome, a full understanding will require work on the non-coding variants, the regulatory regions, and mRNA, and issues such as pleiotropy and sexual dimorphism will need to be elucidated.

Second, there is a need for formal linkages between model organism communities and databases with precision medicine initiatives, Lloyd continued. Such linkages would facilitate cross-talk and improve ontologies and comparative phenotyping. Third, and relatedly, there is a need to enhance processes and expand availability of tools to apply comparative ontologies and develop computable phenotypes. These processes and tools will improve the ability to align phenotypes, mechanisms, and genotypes, and will improve the rigor, reproducibility, and reliability of research, said Lloyd.

Lloyd’s fourth “call to action” is to integrate animal ontologies with in vitro analyses, humans-on-a-chip, and patient data to identify knowledge gaps that propel precision modeling of specific human variants. The integration of these different models and data sources can accelerate the process of distinguishing correlated and causal factors, and can help researchers optimize the selection of model organisms. The thoughtful integration of models may reduce the need for animal models and may improve decision making by creating better evidence with less noise.

Fifth, there is a need to promote, incentivize, and reward a culture of synergistic, clinical outcomes-focused networks and collaborations between translational researchers, clinician scientists, and patient groups, said Lloyd. Collaborations between these sometimes-siloed groups can inform and improve preclinical and clinical trials, can help realize the “regulatory dream” that Califf had presented, and can keep the focus of research centered on patients and health outcomes.

As researchers continue to identify variants that are potentially relevant to human disease or health, we need to expand model organism resources and technology development for rapid assessment of variants, said Lloyd. These resources and technologies may include things like model-on-demand programs, expanded biorepositories and tissue/specimen banks, improved phenotypic validation, and steps to ensure predictive validity of models.

The seventh and eighth “calls to action” that Lloyd presented were to establish clinical utility of a precision modeling paradigm to advance and accelerate diagnostic decision making, targeted therapeutic development, and predictable disease prevention strategies and to ensure accessibility of resources, tools, and networks of precision modelers to the global community. Ensuring accessibility to the global community will help “transcend cultural and socioeconomic divisions, reduce and overcomes disparities, increase access, and increase opportunities,” Lloyd concluded.

Through the eight sessions of the workshop, speakers identified specific challenges to the development of predictable, reproducible, and reliable animal models to support rapid and accurate identification and prioritization of causative factors of disease to support precision medicine research.

Model organisms can improve precision medicine outcomes, provided that formal collaborations are established between researchers using animal models and precision medicine initiatives so that processes and projects such as comparative phenotyping, ontologies and functional annotation of the genome can move forward in tandem. Precision medicine has been embraced on a global scale, thus global collaboration would help ensure shared knowledge and resources.