Using biointelligence to search the cancer genome: an epistemological perspective on knowledge recovery strategies to enable precision medical genomics

Abstract

Genomic profiling is beginning to extend beyond the many applications in discovery research toward direct medical applications that hold the promise of more precise and individualized health-care delivery. There are many barriers and challenges that still need to be overcome before ‘Precision Medical Genomics’ can deliver the promise of more informed patient care, not the least of which is the unmet need for a new conceptual framework for recovering, understanding and translating potentially useful information from a single genome. Although a wide spectrum of scientific strategies, bioinformatic approaches, IT tools and knowledge resources have been developed to support discovery research, the interpretive requirements for recovering clinically useful insights from an individual's genome are different in many ways from those of traditional research goals. In this study, we compare and contrast the fundamental conceptual differences that distinguish ‘research’ to discover generalized knowledge from ‘search’ to recover individualized knowledge. We also consider the merits of applying evidence-based medicine and traditional scientific methods when n=1, and consider an alternative perspective based on a translational engineering approach and intelligence for interpreting genomic information from an individual case. Although the general idea of biological intelligence-based knowledge recovery that we introduce here can be broadly applied for personal genomics across many indications in medicine, we make a case that the need for adopting such a paradigm is greatest for supporting the management of complex diseases, and particularly suited for supporting therapeutic decisions in medical oncology. Early concepts for designing and implementing this kind of ‘BioIntelligence’ solution will be discussed. We also review the anticipated challenges of implementing genomic analysis and biological intelligence-based solutions in the practice of medical oncology by discussing some of the related pragmatic considerations for deploying the first generation of a ‘Precision Medical Genomics’ solution that can evolve and improve over time.