Microsoft Research delves into many areas looking for ways to apply machine intelligence in new and innovative ways. Identifying images, learning and translating languages, predicting the future–all are fair game when it comes to applying the ability for machines to analyze and interpret various forms of data. Today, Microsoft lets us know about another application, this time in the field of molecular biology.
Currently, biology researchers us a gene editing tool call CRISPR to conduct research on how to shut off genes to accomplish various biological objectives. Predicting how editing a gene will impact an organism is key to understanding disease and looking for potential cures, and the complexity of the process results in tremendous computational loads.
Microsoft researchers, in conjunction with the Broad Institute of MIT and Harvard, are working to add a new system into the mix. Titled Azimuth, the new machine learning tool looks at training data and learns to make predictions on data it hasn’t yet evaluated.
Other computer scientists have tried to apply machine learning to CRISPR. Fusi said this project uses a more sophisticated machine learning model than previous efforts, and it also takes into account what worked and what didn’t with the previous models.
“Our goal was to not only understand why some features were important, but also to comprehensively evaluate all the other work that had been done before,” (Microsoft Researcher Nicolo Fusi) said.
CRISPR, or clustered regularly interspaced short palindromic repeats, basically means precisely editing the DNA in living cells, and involves identifying then targeting thousands of places within a gene that might result in breakthroughs allowing, say, the eradication of malaria. Azimuth will streamline the process, allowing researchers to more efficiently and effectively utilize CRISPR.
Listgarten and Fusi, who work out of Microsoft Research’s Cambridge, Massachusetts, lab, said scientists can use their models to figure out the best approach to take to shut down a gene.
The researchers are continuing their collaboration with a predictive analysis project that also will make it easier for researchers to figure out when and where the use of CRISPR to edit one gene will have unintended consequences elsewhere in the genome. Researchers call this an “off-target” effect and it’s one of the biggest hurdles to using CRISPR for things like curing diseases in humans.
Clearly, this is the sort of research that may not pay off for years. At the same time, it’s research that’s vital to continued advancements in medicine, and we’re glad to see Microsoft bringing their machine intelligence chops to the lab. Feats such as curing blindness and creating resilient crops to help feed hungry people are well worth the effort.
Further reading: Machine Intelligence, Medicine, Microsoft, Microsoft Research
