Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Nod Profile picture
Nod
@nizzaneela
Cutter of chrysoprase
Mar 17, 2024 • 11 tweets • 4 min read
"The molecular epidemiology of multiple zoonotic origins of SARS-CoV-2" includes a striking figure that contrasts likelihoods of viral tree shapes consistent with a single introduction (centre and right) vs that of a viral tree shape consistent with two introductions (left). Image Unfortunately, the likelihoods are distorted by errors and noise. We can correct these errors and resample to suppress noise (). This changes the picture slightly, but it still isn't useful. The likelihood on the left isn't really what we're interested in. github.com/nizzaneela/mul…
Image
Feb 20, 2024 • 4 tweets • 1 min read
I used to complain about his models, but now that I can reproduce them I'm more worried about how four of his five analysis cases failed to produce a single CC topology. I'm serious. For the primary analysis, I get a 1/10 likelihood from 1000 complete resamples of the final stochastic phase (mutations generated by "stableCoalescence_cladeAnalysis.py") using the 1100 time trees from the Zenodo data. I'm worried the other four analyses... Image
Oct 25, 2023 • 9 tweets • 3 min read
In case anyone is interested, I can provide a little more detail on how we found these errors. A lot of people were skeptical of the modelling behind Pekar '22. I was particularly worried about the way epidemics were simulated over simple networks that lacked realistic structures and instead had a scale-free degree distribution.
Nov 8, 2022 • 19 tweets • 6 min read
There have been some recent discussion about bad faith arguments in e-letters to Science. Some especially disappointing examples can be found under Pekar et al.‘s “Timing the SARS-CoV-2 index case…” it 2021 science.org/doi/10.1126/sc… The authors used BEAST to infer a probability distribution for the time of the most recent common ancestor (tMRCA) to the virus samples taken from China between December 2019 and April 2020.
Mar 31, 2022 • 24 tweets • 8 min read
This thread is for substantiating the argument that the polytomy in the simulations of Pekar et. al is an artefact of the model specification. The polytomy underpins the low probability ascribed to lineages A and B diverging from a single spillover. Image