Hayden Metsky
I lead research and product development at Valthos, a startup building AI systems that identify biological threats and design medical countermeasures in real time. I live in New York City.
Previously, I worked at Inceptive, designing RNA-based medicines using AI and high-throughput experiments. As a group leader at the Broad Institute and a graduate student researcher in the Sabeti Lab, I developed genomic tools to detect and track pathogens at scale.
I earned my PhD in computer science from MIT, where I also completed my SB in computer science and in physics.
Papers
My full publication list is on Google Scholar. Below are papers I'm most proud of.
Designing sensitive viral diagnostics with machine learning
Capturing sequence diversity in metagenomes with comprehensive and scalable probe design
Zika virus evolution and spread in the Americas
Video Broad Institute news Nature · News & Views Time The Atlantic Wired NBC
Software
ADAPT
ADAPT designs highly sensitive and specific viral diagnostics, with a focus on CRISPR-Cas13–based diagnostics. ADAPT employs a deep learning algorithm to devise diagnostic sequences with maximal sensitivity across the full spectrum of a virus's genomic variation. The approach takes into account all publicly available genome sequences for each virus. ADAPT enables a resource of highly sensitive, species-specific CRISPR-Cas13–based diagnostic designs for 1,900 viral species that stay automatically updated to reflect the latest viral variation.
Design app and resource of designs for >1,000 viruses GitHub Related repositories COVID-19 designs
CATCH
CATCH designs probe sets for targeted enrichment of diverse genome sequences, enabling more sensitive and cost-effective metagenomic sequencing. The probe sets, which comprehensively and compactly reflect extensive genomic variation, improve detection and genome characterization of viruses and other microbes from samples with unknown contents. CATCH employs alignment-free algorithmic techniques to cope with vast sequence diversity. We have used CATCH to design panels that successfully enrich whole genomes of all human-associated viral species and all their known subspecies variation. It has also been applied by groups to generate targeted human viral surveillance panels, as well as agricultural viral, bacterial, and human leukocyte antigen (HLA) typing panels.