Hacker News Reader

Perhaps, a better way to traverse this would be to look at the history of virology, trace main events, starting from the discovery of bacteriophage, and check out associated work. It is hard to say what kind of evidence would satisfy you. Viruses have been isolated many times, and transmissibility has been shown on cell cultures and animal models.

You can't purposely infect real people with disease-causing viruses in research so you won't find much studies like that. But there is no doubt that viruses are transmissible and can cause disease, and there is plenty of evidence to support that. Check out studies on yellow fever virus and flu done more than a hundred years ago.

If you read the Wikipedia page on Koch's postulates you'll see that Koch himself abandoned and modified them for various bacteria. https://en.wikipedia.org/wiki/Koch's_postulates#Postulates

Viruses, and some unculturable bacteria and other organisms, are simply unique with regard to Koch's second postulate, in that they literally can't be cultured in a pure manner. Hierarchical and tangential flow filtration can, however, get them incredibly pure (as quality controlled by electron microscope) after culturing. I'm not an expert here. This is probably not routinely done as it would be costly and to no purpose for the particular experiment. But it can be done, and I'd be extraordinarily surprised it it hadn't been done.

> We've never, again as far as I'm aware, isolated a virus from a sick person, exposed a healthy person to it, and seen the same disease symptoms show up with the virus now present in their system.

Yes, we also don't typically expose healthy people to disease-causing bacteria either. This is a hard sell post-WWII.

Typically when you culture a virus you'll also have a separate culture that is exposed to everything else minus the virus. You'll "purify" both in the same manner, expose some animals or cultured cells to the purified virus culture, and expose other animals or cultured cells to the purified control culture. The only thing that differs is the presence of the virus.

You can also literally view cells being infected by physical viruses in culture, and image or video what happens to them during this process. E.g. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6330349/

Read the "Routine Diagnosis" part of this article.

> > For the more straightforward diseases it's even relatively straightforward to show that a particular virus causes that particular disease, with enough work.

> I'd be really interested what this work would look like.

A stringent example:

1) Replicate the virus in cell culture and ensure you're isolating the virus while doing so.

- A: homogenize and passage a virus containing sample.

- B: passage it through cell culture.

- C: dilute the supernatant from cell culture until the dilution, on average, has less than one virus per aliquot.

- D: Use these aliquots to inoculate further cell culture.

2) Analyze all cultures

- A: As, on average, some of the aliquots contained no virus, there should be a number of cultures which look identical to the uninfected controls.

- B: Any cultures which show morphology or pathology distinct from the uninfected controls should contain specific virus particles and show specific viral RNA/DNA on a PCR test. Likewise, any cultures which look like the uninfected controls should be free of virus based on these tests.

- C: If any cultures show morphology or pathology distinct from the uninfected controls but do not contain virus particles (given sufficient time for the virus to propagate and emerge from the cells), then you've got a thinker on your hands. Triage for contaminants or another disease causing agent.

3) After demonstrating that virus is the causal agent of the infected culture morphology or pathology, use multiple of these infected cultures to infect animals (and have control "infections" from the uninfected cultures). Then repeat this process from the infected animals to demonstrate that it is indeed the identified virus which causes the disease.

If you can't use an animal model for the disease then you're stuck doing this in cell culture and comparing the infected culture cell morphology and pathology to that of cells in biopsies of diseased people.

Pretty darn good evidence that it is a virus, and a particular virus at that, which causes the disease.

Realistically it's much more convenient just to use 2B as evidence. You get some issues with natural immunity and lysogenic versus lytic viral stages (these have different terms in Eukaryotic viral infection), but it's pretty decent for showing things like HIV being the causative factor of AIDS even prior to showing the effect of HIV presence in immune cells.