Harris et al posted a paper in October called “On the unfounded enthusiasm for soft sweeps II,” in which they presented an alternative explanation for a pattern I observed in my first PhD chapter. https://www.biorxiv.org/content/early/2018/10/19/443051 … 2/
-
-
Prikaži ovu nit
-
Some background: In 2014, Pennings, Kryazhimskiy and Wakeley published clear, haplotype-based evidence of soft sweeps in response to poor therapies and speculated sweep softness might be related to therapy efficacy. https://doi.org/10.1371/journal.pgen.1004000 … 3/
Prikaži ovu nit -
We followed up and found that early in the HIV epidemic, when drugs failed often due to drug resistance, HIV populations fixed mutations without a loss of diversity. As drugs improved, more diversity was lost after fixation of resistance. 4/
Prikaži ovu nit -
We interpreted this as evidence of a decrease in the population mutation rate, and a shift from soft to hard sweeps. We published this in 2016: https://elifesciences.org/articles/10670 5/
Prikaži ovu nit -
Harris et al claim higher post-sweep diversity after early therapies could be generated if sweeps to bad drugs were slower and followed larger population bottlenecks, while sweeps to good drugs were faster and followed more severe bottlenecks. 6/
Prikaži ovu nit -
Then, recombination could rescue more diversity during sweeps to bad drugs than good ones. Here’s a comparison of our models: 7/pic.twitter.com/NrRU2hPwTB
Prikaži ovu nit -
This is a totally valid hypothesis. It’s one we considered when writing the paper, and I mention it when I give talks about the work. I don’t think it explains diversity maintenance in HIV under bad drugs. 8/
Prikaži ovu nit -
It turns out that HIV fixes mutations in response to bad therapies really quickly and predictably. Here’s some data from ~20 patients treated with FTC in the early 1990s: 9/pic.twitter.com/LBPbjraz9u
Prikaži ovu nit -
In order for resistance to increase so *quickly*, s must be large. When selection is so fast, recombination has little time to operate. 10/pic.twitter.com/DCZHFz1aAj
Prikaži ovu nit -
In order for resistance to increase so *predictably*, theta must also be large. If the probability of a sweep across populations is so high (as it was early in the HIV epidemic), those sweeps are probably soft. 11/pic.twitter.com/UzOzFxOmy9
Prikaži ovu nit -
In summary, Harris et al present a reasonable hypothesis, but it doesn’t fit within the constraints of the HIV system. I think the authors missed some important parts of the HIV literature. 12/
Prikaži ovu nit -
It’s not fun to have your work criticized, but I’m trying to look at the positives of writing the paper: 13/
Prikaži ovu nit -
I had an excuse to try out SLiM (it rocks,
@MesserPhilipp + Ben Haller). I read some cool new (well, old) papers. I learned about Rapoport’s rules: https://rationalwiki.org/wiki/Rapoport%27s_Rules … (h/t Noah Rosenberg). 14/Prikaži ovu nit -
I also had many valuable conversations with colleagues (esp officemate
@benjaminhgood and co-authors@pleunipennings and@PetrovADmitri), and I feel like I have a better understanding of the system than when we wrote the original paper. 15/15Prikaži ovu nit
Kraj razgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.