If you're designing a new protein, what matters is that it gets used. It can be sold as a product, it can be probed for understanding or it can be used to examine another system.
If your protein is going to be used in one of these ways, then it needs to be manufactured. If you're selling this protein as a product, then to make it economically viable it will need to be manufactured at a medium-to-large scale. Can you do this? What PTMs, expression system, and stability requirements does your protein have? How will this dictate manufacturing constraints? With those constraints, can you sell enough of it to make money?
I'm writing this because this video was retweeted recently and it made me reflect on the headlines I see, the work I've pursued, the hype I've bought into in the field:
It's easy to confuse the essence of what you're doing with the tools that you use.
Scientists, protein engineers, programmers - all of us love our tools. I love tools, I love making them, optimizing them, tweaking them to no end. Cool tools get funded. But at the end of the day they should be used to make something.
There is so much unsexy, boring work to be done in this field. There are hundreds of millions of proteins we can make. So many useful datasets to be made. The big models are cool, new algorithms are essential, but I hope to see more boring work that matters.