if you ever wondered how the lesbian lizards¹ maintains genetic diversity: it has twice the 2n chromosomes & recombines sister ones. SciAm calls it² asexual but it would be more accurately described as selfing-only hermaphrodite from genetic point of view ¹en.wikipedia.org/wiki/Desert_g

The lizards are dependent on being roughly fuc^W^Wcopulatory behavior for ovulation³, so they screw each other⁴ take turns being a top and a bottom³,⁴

@munin the technical term is 'neuroendocrine primer', yes


@munin yes; it's unusual but not to the degree it was hyped up to. do you know how your immune system generates antibodies? T/B cells rearrange and mutate their own DNA, then any ones that result in anti-self antibodies are eliminated, and then the ones that result in ineffective antibodies are outcompeted. You get something like 2B possible antibodies

@whitequark ....Ok, that I did not know, and that is -incredibly- cool.

@whitequark Wait, is this why autoimmune stuff happens? Because the anti-self-elimination messes up?

@munin sometimes! Also blood-brain and especially blood-testicular barriers make it harder for immune system to surveil those organs, hence if you get a vasectomy you start making antigens for your own sperm, which is likely to make you infertile.

@munin Immune surveillance is like... how does immune system find out about intracellular viruses or parasites? Well everything that the cell degrades, it breaks up in pieces and then sticks on its outer surface, so that B cells (iirc) can lick it and see if it got infected, ie presenting non-self antigens

@munin In the brain you actually have a separate blood-cerebrospinal fluid barrier, which filters plasma destined to become CSF, and cells of that barrier *also* look at the crap that floats in CSF and present it back to the immune system by putting it on their blood-facing surface

@whitequark ....huh. Kind of like having a separate network partition for the backhaul [heh!] infrastructure.

@munin The self/non-self distinction is made by choosing (early into development) a set of antigens with a CSPRNG^Wreally good randomization that's called "major histocompatibility complex" and different for every organism, T cells are trained to not bind to self MHC (tissue allorecognition)


....wait...my immune system depends on what amounts to a bootstrap keygen process?

@munin precisely. This is why xenograft (transplant) rejection is a thing. MHC is similar in siblings etc and same in identical twins so those do not reject transplants (in case of identical twins those are allografts, same as if you donated your own organs before)

@munin What's more, MHC class I is *also* used for antigen presentation I just talked above, but even better! There's MHC class II, expressed by professional antigen-presenting cells, such as B memory cells. They suck up antigens and keep the immune system trained on them, often for many decades or life

@munin Correct. You don't need the entire bacterium or virion, you just need characteristic parts of it so your immune system can get specifically trained to recognize it

@munin Some pathogens, like Plasmodium falciparum or Mycobacterium tuberculosis, interfere with this process; P.f. just ensures that it isn't digested, and M.t. actually downregulates MHC class II synthesis in antigen-presenting cells to avoid detection

@munin There's more! You also have an innate immune system; bacteria synthesize their proteins starting with N-formyl-methionine so you have receptors (Toll-like receptors for that one) for this pattern and hundreds of others that trigger inflammation and so on

@munin You might have heard that mitochondria and chloroplasts are descendants of cyanobacteria; logically, they'd start protein synthesis with N-formyl-methionine. Well they do! So if your muscle breaks down then you get inflammation because your own mitochondria look sufficiently like bacteria

@munin The immune system is incredible from an infosec POV

@whitequark It reminds me a lot of how suricata/snort directives are written, and how the $HOME_NET vs. $EXTERNAL_NET type decisions are made

@munin It amazes me that this shit can adapt to a virtually unlimited set of chemical compounds. You know how we e.g. measure estradiol levels in labs, or drug test athletes, etc? We inject whichever thing we want to analyze into a rabbit or w/e, that forces the rabbit to make an antibody to it. Then, we isolate the B or T cell expressing the antibody, and merge it with a cancerous cell so it reproduces (and produces the antibody) forever

@munin Then we isolate the actual antibody chemically and fix known amounts of it on a tiny test tube. Put the blood (usually plasma) sample into it, hold it there, wash, then analyze how much got bound (there are a few methods). This is called "immunoassay" and is an INCREDIBLY popular and often only method for precise picogram analysis

@munin You can buy hundreds of thousands of pre-made antibodies eg abcam.com/kits/simplestep-elis pretty cheaply and then analyze stuff with basic tools (just the IR spectrometer past the basic lab tools, maybe a centrifuge to remove blood cells if you use blood)


...this is pretty amazing. So it's like custom reagent production, kinda.

@whitequark .....! Oh good grief, that explains so much about muscle injuries

@munin "damage-associated molecular patterns" is the keyword fwiw

@whitequark ....suddenly a lot of things make a lot more sense. This is pretty awesome.

Sign in to participate in the conversation

Server run by the main developers of the project 🐘 It is not focused on any particular niche interest - everyone is welcome as long as you follow our code of conduct!