Research Stories

VP(R)'s Picks: Faster, Better Tool Advances Precision Medicine

Vice-President (Research)'s Picks: Faster, Better Tool Advances Precision Medicine


Related Links

Precision medicine offers hope of better diagnoses and treatments for patients. But to achieve that, doctors need to sift through the three billion characters in each person’s DNA. Groundbreaking research by Dr Luo Ruibang, published in Nature Communications, is making that task a lot easier.

Dr Luo Ruibang has developed a new tool that can extract valuable genetic information from noisy backgrounds in a way that was previously impossible.

Clairvoyante is an open-source deep learning model that uses artificial intelligence to accurately cull weak signals of genetic variants, or mutations, from the noise in single-molecule sequencing, using only a laptop computer. It offers better precision and recall than other state-of-the-art models and it can find variants in less than two hours – two orders of magnitude more time-efficient than the existing methods.

The finding opens the possibility that doctors may be able to do accurate genetic variant detection with just a hand-held sequencing device and a laptop, enabling molecular diagnosis to be done in remote and less-developed regions.

The discovery is also effective at detecting pathogens, such as viruses, and Dr Luo is developing this research path with Johns Hopkins University.

Dr Luo previously developed an algorithm that has become a global standard for genome assembly because it reduces the process of reconstructing a genome billions of nucleotides in size from about one month to less than a day. He also developed a big data analytics platform that assists doctors in diagnosing cancers and rare diseases by quickly comparing a person’s genetic data against multiple clinical databases.

Dr Luo was named one of the top 10 innovators in Asia Pacific in 2019 by MIT Technology Review, and one of the top “30 Under 30” in Asia in 2017 in healthcare and science by Forbes.

 Clairvoyante network architecture and layer details.

Clairvoyante network architecture and layer details.

(Image reproduced/adapted from Springer Nature under CC BY 4.0: Luo R, Sedlazeck F.J., Lam T.-W. and Schatz M.C., “A multi-task convolutional deep neural network for variant calling in single molecule sequencing”, Nature Communications, 2019, 10, 998, 1-7.)