The Berlin Zoo named their baby Hippo Brötchen. :) (Bread roll 🥖)
Video
Knoebel
4/ One is the exponent of the SFS decay, a proxy for growth (1 = constant size, 2 = exponential). The other is the number of clonal mutations. Immunocompetent tumors land on fewer clonal mutations, consistent with immune restriction.
6/ Against that null, total binding neoantigens are as expected, but the clonal ones in immunocompetent tumors are stripped of strong binders. We're actively improving the ABC side now, so feedback welcome.
1/ New paper with Elizabeth Borden, Karen Hastings, and Ryan Gutenkunst on immunoediting in human tumors. Ryan and I built the modeling side, so that's what I'll focus on here.
2/ The setup: skin SCC from immunosuppressed transplant patients (immune brakes off) vs immunocompetent patients. Same mutation load, but immunocompetent tumors have far fewer clonal mutations. Is that the immune system, or just growth dynamics?
5/ Our second piece is a null model: the neoantigens you'd expect from a tumor's mutation pattern alone, no immune system. We generate in silico mutations from the trinucleotide spectrum and per-gene mutability, then predict MHC-I binding.
3/ Our first piece is an ABC method. We'd love to separate growth from immune pressure directly, but can't. Instead we treat each tumor's variant allele frequencies like a noisy site frequency spectrum and infer two things from it.
7/ Huge credit to Karen and Elizabeth, who led this. I learned a ton from both of them. Elizabeth especially is a remarkably talented early-career scientist, very much one to watch. Paper: url.usb.m.mimecastprotect.com/s/XxjOCnGWL0...
Of course the link does not work! www.sciencedirect.com/science/arti... @ryangutenkunst.bsky.social
