Here is a new paper from the Regeneron Genetics Center A new, fast ML method for whole genome regression. The REGENIE website is here with free source code. Amazing work from 1/10
The REGENIE method has an Apache Spark implementation in GLOW, which allows even better compute scaling. Fitting the whole genome regression step take minutes. See this nice blog for more details
Apparently this tweet from Jonathan has been censored due to its 'sensitive content' 👇 Not surprising as it uses to show how much CO2 a GWAS analysis could save by switching to Hint: It's a lot.
We used to compare the carbon footprint of REGENIE to SAIGE in terms of carbon emissions, carbon sequestration and the distance a car would travel to emit the same amount of CO2 used by the methods