How do we actually ever figure out whose neoantigen predictions are right? And is there anything that your project can do about that?

So this is the N equals one problem.

Paul also knows N equals one opportunity because each drug is designed for patients.

So how do you know that it works? And so people scratch their heads for a long time about this.

And some people say, actually, you should just give it to people with rare diseases.

And I think for rare pediatric diseases, people just want to try and get these drugs out there and will assess whether the children survive or not.

And I think that's a very, very fair way to do that.

i think we have seen in so many actually distant fields that the real power of ai is when there's a closed loop between what's happening in in silica prediction and then a real world test of that and then you put those failures and successes back into the algorithm and you refine it so i believe that the measure of t cell response is done through ellispot analysis and maybe you could just tell us briefly about how you're going to automate and scale the LSBOT analysis in your program to just figure out which neoantigen prioritization pipeline is most likely to put the right antigens and the right MHC peptide complexes to drive a CB4 or CB8

Mm-hmm.

That's really special.

Um, and also, she's increased coverage to, uh, did you say sixty percent of the population? That's forty-two million people have potential access to this, so that is huge.

And I think it's really a passion project for so many people out there, p-trials units, the research nurses and doctors, and also patients who make this all happen.

So it is quite impressive.

It is very impressive.

Oh, yes, and AI.

Yeah.