This current project is effectively a cache database for an earlier tool. The earlier project is where the main impact is - predicting the functional effect of a variant in the human genome.
For this updated project, they effectively ran the earlier tool for all possible variants in the genome (3 billion bases in the genome * 3 single nucleotide variants = 9 billion). It’s a very useful project for people like me, but the major advance was already done (admittedly by the DeepMind team).
There are a lot of companies and academic institutions working on genomics. Most don’t do self promotion at the scale that Google does. And most aren’t on the HN radar.
7 comments:
Prior post: https://news.ycombinator.com/item?id=49611251
google PR machine on relatively low-impact science goin pretty hard
Do you really think that tho? I mean, it is the only one doing something other than all-in LLM BenchMaxxing lmao :D
This current project is effectively a cache database for an earlier tool. The earlier project is where the main impact is - predicting the functional effect of a variant in the human genome.
For this updated project, they effectively ran the earlier tool for all possible variants in the genome (3 billion bases in the genome * 3 single nucleotide variants = 9 billion). It’s a very useful project for people like me, but the major advance was already done (admittedly by the DeepMind team).
There are a lot of companies and academic institutions working on genomics. Most don’t do self promotion at the scale that Google does. And most aren’t on the HN radar.
Now do the combinations of them.
boom