This is the blog for GW students taking Human Evolutionary Genetics. This site is for posting interesting tidbits on: the patterns and processes of human genetic variation;human origins and migration; molecular adaptations to environment, lifestyle and disease; ancient and forensic DNA analyses; and genealogical reconstructions.

GWHEG figure

GWHEG figure

Monday, January 29, 2018

Google's DeepVariant provides new perspective in genome sequencing

Image cred: Ben Bours (copied from Wired article)

Many new genome sequencing techniques have been introduced since the Human Genome Project in 1987. Today, even the tech giant Google has stake in the game, having released their own genome sequencing tool, DeepVariant, late last year. The tool uses machine learning to identify inherited mutations within an individual's genome. According to a 2016 paper by Poplin et al., DeepVariant works by using AI similar to image recognition software to spot variants in a sequence when compared to a reference genome. The new tool has also been shown to surpass other sequencing techniques in efficiency and accuracy. Given the success of DeepVariant, it has been rumored that other tech companies such as Amazon may also enter the genome sequencing market, further integrating consumer tech with bioinformatics and healthcare. DeepVariant embodies a breakthrough in genome sequencing as we know it, and may have implications for how our genetic information is used in conjunction with consumer technology in the future.

To learn more, here is a link to the Wired article and also the original article published in bioRxiv.

Angie Peña

1 comment:

  1. An interesting rebuttal of the Wired article from a professor at Johns Hopkins, in Forbes:

    https://www.forbes.com/sites/stevensalzberg/2017/12/11/no-googles-new-ai-cant-build-your-genome-sequence/#5399c03d5774

    ReplyDelete