My thesis in the Steinmetz Lab involved using computational tools to find RNA secondary structures and tertiary topologies that allow viruses to assemble around their RNA genomes. Even with the secondary structures mapped, it is difficult to determine how to assemble viruses and virus-like particles in vitro.
If you look at a phage or a virus infecting a cell, how does their capsid decide which RNAs to encapsulate? How do we not end up with empty capsids or ones containing host RNA? Secondary and tertiary structures.
I was able to isolate the minimal stem loop needed for the origin of assembly -- about 50bp -- from potato virus x and added it to a GFP mRNA. Assembly. Okay how about the nearest relative of PVX? I tried this with mycovoruses, phages, plant viruses etc. And it works as long as they are filamentous or rod-shaped viruses. I even swapped stem loops around and saw bidirectional assembly (paper almost published). I added it to circRNA and made the first halo-shaped virus-like particles that don't exist in nature. Possibilities of this are endless for therapies, agriculture, phage therapy, and so much more.
I think we have only touched on a few use cases for controlling RNA shapes - still a lot to discover
My rotation project in the Church lab was on structure prediction for bacterial T-box riboswitches (the goal was to classify which tRNA was recognized by the riboswitch; fortunately we only needed to predict the secondary structure for this goal). Even these were rather difficult to model. We looked into deep learning methods but ended up going with a HMM approach due to lack of structural data. And of course, these are ribo *switches*, which change structure upon ligand binding.
So it's very cool to see the progress that has been made since 2020.
The meta-question here seems to be something like: "this scientific problem is very hard, it is probably not useless to solve but also isn't guaranteed to be incredibly useful to solve, how much is it worth attacking?"
I think this same issue confronts a lot of fields in science. I did my PhD mainly in organic reaction mechanism, which is also not useless but not clearly a game-changer for any particular application. Nuclear shell structure calculations are similar—as I've been told, there used to be a lot of interest in the field, but it turns out that understanding the quantum structure of nuclei just... doesn't really matter for anything super important?
My takeaway here is that problem selection is incredibly important for impact-minded people, and that it's easy to get nerd-sniped by something hard and interesting that doesn't have an impact proportional to its difficulty. (Note: not a claim about RNA structure prediction specifically!)
My thesis in the Steinmetz Lab involved using computational tools to find RNA secondary structures and tertiary topologies that allow viruses to assemble around their RNA genomes. Even with the secondary structures mapped, it is difficult to determine how to assemble viruses and virus-like particles in vitro.
If you look at a phage or a virus infecting a cell, how does their capsid decide which RNAs to encapsulate? How do we not end up with empty capsids or ones containing host RNA? Secondary and tertiary structures.
I was able to isolate the minimal stem loop needed for the origin of assembly -- about 50bp -- from potato virus x and added it to a GFP mRNA. Assembly. Okay how about the nearest relative of PVX? I tried this with mycovoruses, phages, plant viruses etc. And it works as long as they are filamentous or rod-shaped viruses. I even swapped stem loops around and saw bidirectional assembly (paper almost published). I added it to circRNA and made the first halo-shaped virus-like particles that don't exist in nature. Possibilities of this are endless for therapies, agriculture, phage therapy, and so much more.
I think we have only touched on a few use cases for controlling RNA shapes - still a lot to discover
My rotation project in the Church lab was on structure prediction for bacterial T-box riboswitches (the goal was to classify which tRNA was recognized by the riboswitch; fortunately we only needed to predict the secondary structure for this goal). Even these were rather difficult to model. We looked into deep learning methods but ended up going with a HMM approach due to lack of structural data. And of course, these are ribo *switches*, which change structure upon ligand binding.
So it's very cool to see the progress that has been made since 2020.
Amongst everyone else, I always look forwards to your comments the most
The meta-question here seems to be something like: "this scientific problem is very hard, it is probably not useless to solve but also isn't guaranteed to be incredibly useful to solve, how much is it worth attacking?"
I think this same issue confronts a lot of fields in science. I did my PhD mainly in organic reaction mechanism, which is also not useless but not clearly a game-changer for any particular application. Nuclear shell structure calculations are similar—as I've been told, there used to be a lot of interest in the field, but it turns out that understanding the quantum structure of nuclei just... doesn't really matter for anything super important?
My takeaway here is that problem selection is incredibly important for impact-minded people, and that it's easy to get nerd-sniped by something hard and interesting that doesn't have an impact proportional to its difficulty. (Note: not a claim about RNA structure prediction specifically!)