Related: Making Vaccine
In my limited understanding of the workings of the open-source radvac vaccine, it is surprisingly basic. In particular, if the peptide design works, then simply ordering and snorting it is sufficient. Most of the rest of the process, which is about creating nanoparticles to which the peptides can attach, just serves to boost effectiveness.
When I say this is "basic", I still want to commend the radvac team and johnswentworth for the combined ingenuity and agency required for making it happen. Sometimes it takes true skill to find a simple solution.
Especially when most people aren't even looking.
I spent most of last year not even considering the hypothesis that it might be possible to make safe, cheap and effective vaccines at home. For a year I allowed my life to be governed by a constraint that a more competent version of me could have relaxed in a few months.
It feels a bit like going to Hogwarts and seeing someone do transfiguration for the first time. The thoughts that should be racing through my mind are: how do I learn such magic? And: if I could master such magic... what else could I use it for?
I'm curious for others' input as I think through that.
What else could the general method of peptides, administered intranasally, be used for?
Other diseases, things other than vaccinations, or even enhancement rather than prevention?
My knowledge of this domain is very limited, and these questions might not have interesting answers. But they certainly seems worth at least posing, and I am curious about the result.
I went down the neoantigen rabbithole, and it was quite interesting.
I liked this talk on "Developing Personalized Neoantigen-Based Cancer Vaccines".
It seems a core part of their methodology is using machine learning to predict which peptides will elicit a T-cell response, based on sequencing the patient's tumour. (Discussed starting from around 11 minutes in.)
They use this algorithm, which seems to be a neural network with a single hidden layer just ~60 neurons wide, and some amount of handcrafting of input features (based on papers from 2003 and 200... (read more)