Figure: Cost per human genome, 2001-2022 (source)
For starters, tools for interpreting the vast amount of information in the human genome have not evolved as rapidly. A typical WGS data set might consume around 100 GB of disk space, and storage and retrieval are significant issues to contend with. It is incredibly difficult to extract any useful information from a WGS data set using the existing tools available to genomic scientists, let alone consumers. Genomic science is still mostly focused on two percent of the genome that codes for proteins, and the rest (the “non-coding genome”) is largely treated as evolutionary junk — which is certainly not the case. Finally, running GWAS or similar studies on large populations using entire WGS data sets remains a massive undertaking due to its computational complexity [5].
Nevertheless, there are reasons to be optimistic. Comprehensive national initiatives such as UK Biobank are making vast WGS data sets available to qualified researchers. WGS tests are getting cheaper and cheaper: it will soon cost more to develop custom genotype arrays and process them individually compared to using WGSs to provide equivalent information — and much more. Latest advances in generative AI are making it easier to extract useful information and insights from massive biological data sets [6]. Medical foundation models are powering the next generation of clinical decision support tools [7]. One of our ventures, GeneLeap, is using generative AI-based protein models to optimize drug treatment for individuals based on their WGS data [8]. Large genomic data sets are being used to develop increasingly sophisticated knowledge graphs to elucidate relations between genomic biomarkers, physiological mechanisms, and disease conditions [9]. In the US, the FDA is making great strides in bringing AI-based decision support tools into clinical practice [10].
Based on the developments listed above, we believe that it is not a matter of if but rather when WGS data sets will become part of routine clinical practice. A regulatory environment that is open to AI-based innovations in clinical care will pave the way to significant growth opportunities around existing use cases in the use of genomic biomarkers in diagnostics and therapeutics. In addition, a host of new clinical use cases will emerge including drug treatment optimization integrated into routine clinical care, genomics-based drug interaction alerts, better access to suitable patient populations for clinical trials, and timely risk identification for diseases with strong genomic underpinnings. With a comprehensive portfolio of ecosystem companies powered by generative AI, Leap AI is positioning itself to be a major player in the next era of genomic medicine.
[1] Large Study Reveals PTSD Has Strong Genetic Component Like Other Psychiatric Disorders. UC San Diego Health, 2019. https://health.ucsd.edu/news/press-releases/2019-10-08-study-reveals-ptsd-has-strong-genetic-component/
[2] Barin-Le Guellec, Chantal et al. Toxicities associated with chemotherapy regimens containing a fluoropyrimidine: A real-life evaluation in France. European Journal of Cancer, Volume 124, 37 - 46. https://doi.org/10.1016/j.ejca.2019.09.028
[3] The 1000 Genomes Project Consortium. A global reference for human genetic variation. Nature 526, 68–74 (2015). https://doi.org/10.1038/nature15393
[4] Single nucleotide polymorphism (SNP) arrays. WikiFreedom. https://www.freedomgpt.com/wiki/single-nucleotide-polymorphism-snp-arrays
[5] Pan, H., Liu, Z., Ma, J. et al. Genome-wide association study using whole-genome sequencing identifies risk loci for Parkinson’s disease in Chinese population. npj Parkinsons Dis. 9, 22 (2023). https://doi.org/10.1038/s41531-023-00456-6
[6] K. Huang, T. Zeng, S. Koc, et al. Small-cohort GWAS discovery with AI over massive functional genomics knowledge graph medRxiv 2024.12.03.24318375. https://doi.org/10.1101/2024.12.03.24318375
[7] Moor, M., Banerjee, O., Abad, Z.S.H. et al. Foundation models for generalist medical artificial intelligence. Nature 616, 259–265 (2023). https://doi.org/10.1038/s41586-023-05881-4
[8] GeneLeap. https://www.geneleap.com
[9] Chandak, P., Huang, K. & Zitnik, M. Building a knowledge graph to enable precision medicine. Sci Data **10, 67 (2023). https://doi.org/10.1038/s41597-023-01960-3
[10] US Food and Drug Administration. Artificial Intelligence and Machine Learning in Software as a Medical Device, 2025. https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-software-medical-device
.svg)
.svg){kind=icon}
