> Of the seven genes added to this organism for normal cell division, scientists know what only two of them do.
This gives you an appreciation of how little we know. I work in bioinformatics, and every one in the field will tell you that there's been an explosion in the number of datasets. However, try finding recent data on your condition of interest! The number of experiments grows incredibly fast, but we're not there yet to build a comprehensive model even of a simple organism.
I wonder if this is going to be incredibly useful or totally over the top for research. Imagine a development like the BioNTech vaccine. With an atomic-resolution cell simulator, you could verify your mRNA code, including the delivery into the cell itself and watch it perform.
But would that be like watching my CPU decode and fetch instructions, interesting but completely useless for most practical applications, or like a debugger introspecting a running process, an incredible useful tool?
IMO, an atomic-resolution cell simulator would be extremely useful.
If you have an atomic-resolution cell simulator, you can add lots of automatic interpretation to layer up the information. For example, you can use relatively basic inference software to get a list of metabolite interactions (and their rates and derivatives). An additional layer would contain inferences about the way those metabolites move inside the ER and so on.
Even with those layers, the amount of brain power needed to understand anything would be significant. We may not have enough experts or expert-years to accomplish much. Perhaps a further step would be to make the models more accessible to people with less specialization. Even find more playful names for cell metabolites, the ones we use right now are utterly unpronounceable.
> But would that be like watching my CPU decode and fetch instructions, interesting but completely useless for most practical applications, or like a debugger introspecting a running process, an incredible useful tool?
It would be like watching the cloud of electrons moving over transistors across an entire chip, without the benefit of knowing quantum mechanics. Except in this analogy, everyone is in the same boat and working on a single codebase that's the surviving vestige of billions of years of "just good enough" coders fighting it out, so every little bit helps.
so fyi I worked in that lab, we called many of them MUFs ("membrane proteins of unknown functions")... We kind of suspect that all they do is maintain membrane integrity and isotonicity just by being present. One easy way for a cell to do to increase the yield of a stuff is to have literally more genes.
So there are a lot of interesting things to do (which I don't know if they did), like instead of having those last five genes in there, copying one of those five genes five times...
This gives you an appreciation of how little we know. I work in bioinformatics, and every one in the field will tell you that there's been an explosion in the number of datasets. However, try finding recent data on your condition of interest! The number of experiments grows incredibly fast, but we're not there yet to build a comprehensive model even of a simple organism.