(while the underlying biology is very different, there are some striking parallels to the migration of bacterial strains across different local gut microbiomes)

We think these intermediate levels of migration could be very interesting, by preserving the large-scale spatial structure of GCs, while still allowing positively selected lineages to expand across a wider range of GCs than would otherwise be possible - potentially balancing diversity & selection.

We show that these local migrations follow a clock-like process @ a rate of ∼1/50 cell divisions - roughly uniform across lineages & time. Plus, migrant B cells continue to evolve w/in their new germinal centers at similar rates, such that the largest lineages in each GC often originate from another

Understanding this effect really changed how I think about the fates of new mutations in the presence of clonal interference – and leads to a mathematical formalism that I hope will be useful in other contexts as well.

In this regime, invading ecotypes effectively "mortgage" their ecological advantage to increase their initial growth rate. But they eventually pay a price at higher freqs, when their ecological advantage suddenly dissipates. This echoes recent findings by P. Barrat-Charlaix & @neher.io in SI models.

We show that despite this large mutational influx, rapidly evolving pop'ns naturally cluster into a smaller # of distinct “ecotypes”, even when their genetic diversity is much larger. This non-eq analogue of competitive exclusion is driven by a dynamical priority effect that favors resident strains.

Most existing models of evolving ecosystems assume that evolution occurs very slowly, so that the ecosystem can always equilibrate before the next mutation appears. Here we focus on the more empirically relevant case where ecology & evolution act on similar timescales, as often occurs for microbes.

Their adaptive reversion model may provide part of the answer, but strong assumptions still needed to account for long-term optimization of ~90% of all protein coding sites. For me, this remains one of the more interesting puzzles of microbial pop gen, since it's such a common trend across species.

How natural selection manages to optimize such weak fitness costs in the face of all the adaptation & genetic hitchhiking we think is happening on shorter evolutionary timescales is still a major puzzle. Previous work by Paul Torrillo & @contaminatedsci.bsky.social suggests that it's not so easy!

We use this finding to re-examine models of purifying selection & adaptive reversion in human gut bacteria. After correcting for HGT, we show that most protein-coding variants are eliminated ~10x more slowly than previously assumed. Yet they are still reliably purged on 10-100k yr timescales.

Many studies have found that w/in-species dN/dS decays w/ the genetic distance between strains, which is often attributed to natural selection. Here Zhiru shows that a large portion of this trend can be quantitatively explained by the accumulation of horizontally transferred DNA segments over time.

The first is from former PhD student Zhiru Liu @zzzhiru.bsky.social (now in @bengrbm.bsky.social's group @ MSK) examining the long-term patterns of selective constraint – measured by the classical ratio of nonsynonymous to synonymous mutations (dN/dS) – within recombining populations of bacteria.