I'm intrigued by the promise of applying ML to synthetic biology to help make genetics/genomics more predictive, but I can imagine this is a gargantuan task. I'd like to learn more about the challenges that this involves: Both immediate and a bit farther out.<p>I'm not an expert on either topic (my background is in old-fashioned biochemistry and cell biology). However, I'm hoping to EITHER be pointed to useful resources that offer a big-picture discussion of problems and challenges OR learn more from people here who might have relevant insight and/or anecdotes to offer.<p>For example, how do you annotate biological data so it makes sense for ML purposes? If you're building something like an ML engine to predict how genetic changes lead to phenotypic effects, how does this work in practice? Can this be done without understanding the 'grammar' of biology and is this synthetic biology or just a fancy kind of old-school bioengineering (using known biology for new purposes)?