Unfortunately, DC's for all antigens in proportion to their population in the tumor is probably not statistically optimum. So I was wrong. But there is a silver lining.
The chances of a programmed DC finding it's antigen is proportional to the product of the number of DC's programmed for that antigen, and the number of that antigen expressed on the tumor cell membrane.
So, take 3 different antigens on the tumor membrane with the populations: (1, 5, 1). Now match that with the population of DC's created as (1, 5, 1). The chances of a DC hit on the tumor will be proportional to (1 + 25 + 1) = 27.
Compare that to just programming all the DC's as the most common antigen, the second antigen, ie (0, 7, 0). The yield there is (0 + 35 + 0) = 35. So the statistical optimum for killing the tumor for these considerations is just using the most common antigen to program the DC's.
So my argument for programming the DC's in proportion to the antigen population spread being the statistical optimum, which I said was a stretch, was more than a stretch. It was wrong.
However, in addition to the aforementioned (by others) issues of a given antigen disappearing due to mutation, and possible problems recognizing synthetic antigens for programming, there is another glaring problem with targeting a single antigen. Note that using a single antigen was stated as optimum by a self proclaimed expert on another blog. That initiated this thread.
IMUC targeted 7 antigens with ICT-107. And they acknowledged that a significant percentage of patients would not be helped by ICT-107. There was debate as to what the percentage was, but even the number the that IMUC touted was either 25% or 50% of patients that would not be well handled by those 7 antigens.
If 7 antigens is broadly known to not cover every GBM tumor, then why in the world would this supposed expert be pushing the idea of using a single antigen? So that is a huge gaff in the apparent expert's reasoning, making him far less credible. Further, the guy was using vocabulary that he knew 95% of his audience would not understand. I think that usually means the guy is not broadly smart. He is show-boating at the expense of communication. That's not smart. I thought he was saying that DC training with tumor Lysate would likely produce DC's programmed against native (un-mutated) proteins, resulting in auto-immune problems. But he ignored the fact that the immune system is known to not behave that way, and that studies were done on that issue many times by NWBO and many others, and this minor concern was proven a non-issue. Further, he didn't describe this mechanism for bad programming. He threw out a buzz-word / technical term, but with no description whatsoever.
Bottom line... he knows some technical terms, but he doesn't appear to know what he is talking about. Or, of course, he may be shorting the stock and deliberately misleading readers.
Market Data 
Markets 
AI
