3 Sep 2023
Back from travel and I had time to spot lots of interesting science to dump here, maybe too many this week. A lot of bacterial and phage microbial ecology this week, a dash of cancer, antibiotics, and covid. Lots of game theory! I like that the focus is changing week to week. We have:
Deep mutational scanning reveals the functional constraints and evolutionary potential of the influenza A virus PB1 protein
Another week, another Lauring Lab paper. They assay 84% of all substitutions in the AH1N1 flu polymerase catalytic subunit, since it’s commonly involved in host adaptations when jumping species barriers. They compare their experimental data of the fitness effects of individual mutations to observational data from flu genomic surveillance. A lot of interesting lessons here, but I’ll note real mutational landscapes are a lot more constrained than the experimental data shows, in good part determined by mutation accessibility.
https://doi.org/10.1101/2023.08.27.554986
In general, the effects of mutation accessibility seem to be a recurring theme in this paper list as of late:
- Predicting evolutionary outcomes through the probability of accessing sequence variants and the “genotype-accessibility landscape” from week 13 Aug https://www.science.org/doi/full/10.1126/sciadv.ade2903
- Simulations of sequence evolution: how (un)realistic they really are and why, showing that we underestimate the degree to which mutations are equally accessible in our simulations, from week 27 Aug: https://www.biorxiv.org/content/10.1101/2023.07.11.548509v1
A ubiquitous mobile genetic element disarms a bacterial antagonist of the gut microbiota. Translating this paper into game theory terms: a mobile genetic element (GA1) induces B. fragilis individuals to become cheaters by stopping expression of a non-mobile TSS6 used to antagonize other bacteria (GA3), which allows the conjugants to outcompete their non GA1-counterparts. Eventually, that probably leads to an excess of non-fragilis Bacteroides that would normally be kept in check by GA3-expressing fragilis, so I bet the equilibrium swings back towards GA3 expression, ending up in nice stable coexistence (I think the authors don’t show this part).
https://www.biorxiv.org/content/10.1101/2023.08.25.553775v1
The architecture of metabolic networks constrains the evolution of microbial resource hierarchies.
Using a metabolic model of E. coli, the authors create virtual mutants with altered metabolisms and preferences for sugar sources. They find that only 15 of the 5040 possible rank orders of 7 sugar sources are able to evolve. They point to the implications this has for the evolution of coexistence between species based on niche availability from carbon sources.
https://doi.org/10.1093/molbev/msad187
Community ecology of phages on a clonal bacterial host.
Now this is really interesting. Environmental phage isolates don’t outcompete each other when cultured in lab conditions. This is counter to expectation from other bacterial ecology experiments, where reconstituting stable communities in experimental conditions is nontrivial. Goes to show how parasitism introduces fundamentally different rules of the game for ecology and evolution. Love it.
https://doi.org/10.1101/2023.09.01.555216
Real-time identification of epistatic interactions in SARS-CoV-2 from large genome collections.
The authors expand a method using mutual information to identify epistatic interactions between loci in SARS-CoV-2 based on signatures of selection in genomic surveillance data. When looking at 4 million genomes, they found 474 interactions, but got pretty good results with as little as 10K sequences. In fact, they spotted a key innovation with only the first 6 Omicron sequences. This could be really useful for mapping out landscapes in early outbreaks of other variants or pathogens too.
https://www.biorxiv.org/content/10.1101/2023.08.22.554253v1
Accelerated evolution of SARS-CoV-2 in free-ranging white-tailed deer.
I saw this work presented at EEID in May too. SARS-CoV-2 mutates faster through deer (possibly due to higher APOBEC activity) with different selection pressures, nuff said. We need to keep an eye on this for zoonosis from reverse-zoonosis. Don’t pet deer in Ohio. Just don’t pet deer.
https://www.nature.com/articles/s41467-023-40706-y
Leveraging insect-specific viruses to elucidate mosquito population structure and dynamics
This is a review on a cool idea. By looking at the phylogenies of viruses that infect mosquitoes (and only the insects, not arboviruses infecting other hosts), you should be able to infer facts about the size, structure, and dynamics of those vector populations.
https://doi.org/10.1371/journal.ppat.1011588
Drug mode of action and resource constraints modulate antimicrobial resistance evolution.
The authors use an ODE model to explore whether growth inhibition or active killing is more effective at preventing the evolution of resistance to treatment (in a bacterial infection model, but this framework extends to cancer, etc.). They try single-drug, cycling, and combination therapies, and find that if multiple drugs are used, combining both growth inhibition and active killing is the optimal strategy. An interesting implication not discussed by the authors is that if >2 drugs are used, the optimal ratio of inhibition:active killing depends on how easy it is to evolve resistance to each growth inhibition drug: you want to make sure you always have at least one active growth inhibiting drug (if you only have active killing, it’s just not as good), but having >1 inhibition drug offers diminishing returns, while the effects of multiple active killing compounds on top of an inhibiting drug do stack up nicely. The authors alos made an online app for the model: https://oscar-delaney.shinyapps.io/honours/
https://www.biorxiv.org/content/10.1101/2023.08.29.555413
Games and the Treatment Convexity of Cancer.
This is a bit hard to follow if you’re not familiar with the notation already (I wasn’t really), but it’s a really interesting way of thinking about how cancer treatment works. By modeling tumors as a game between fast and slow growing cells that can have different interactions with each other, the authors classify tumors into four different classes of games being played. Treatment can move the tumor from one class into another as you increase dosage, with some non-intuitive behaviors including dosage hysteresis.
https://link.springer.com/article/10.1007/s13235-023-00520-z
Costs of resistance limit the effectiveness of cooperation enforcement.
This seems interesting, and slime molds are such a fun model system. The authors use slime molds to look at the evolutionary effects of “cheating” strains (in this context, mutants who never sacrifice themselves for the sake of spore dispersion, and instead favor spore production only) and cheating-resistant strains. They find that cheating resistance has evolutionary costs, but not so much in traditional “fitness” (resistant strains grow just as well as non-resistant ones) as in evolvability (evolving resistance to cheating, or any one given pressure, makes it difficult to evolve resistance to others). This is a clever and crucial way to rethink adaptive landscapes, and what “fitness” actually means.
https://www.biorxiv.org/content/10.1101/2023.08.29.555259v1
An antibiotic from an uncultured bacterium binds to an immutable target.
Everyone and their mother has dunked on this paper for suggesting a compound is irresistible based on a single-dosed attempt to observe resistance, and yes I do the same, they deserve it. Still, it’s not every day you get a new class of antibiotics, so check it out. Works against Gram+ only—not ideal, but still!
Comments
Post a Comment