Justin Conover

And as always, I'm happy to answer any questions or receive any feedback on this work!

I hope Sam (who is a math/economics major, and had never taken a genetics class before this project) learned as much about polyploids as I learned from him about math. Thanks also to @ryangutenkunst.bsky.social and @barkerms.bsky.social for overseeing this project as well

Overall, I'm super excited about this paper, as it is one of the first pieces of theory based on tetraploids with disomic inheritance, without treating the subgenomes independently. I think there is still lots to learn on this front, and hope that other will help push this field forward

Here, we can directly measure additive genetic variance and the change in allele freq and pop fitness. When the population is in the bistability regime, Fisher's Fundamental Theory of Natural Selection does not hold, and there are subtle differences between disomic and polysomic inheritance patterns

Finally, we wanted to see if there was a difference to how polyploid populations might approach these equilibrium points based on segregation patterns. For this, we developed a forward-in-time 1 locus simulation for disomic and polysomic inheritance.

The ability for a population to remove this is totally dependent on whether the initial allele frequency is higher or lower than the unstable equil point. Also, because this model assumes infinite population sizes, there is no influence of drift in this model.

This is a classic example of a saddle node bifurcation in nonlinear dynamics. Meaning, a population may be completely unable to purge a deleterious mutation, despite it being dominant and at high/intermediate frequency, depending on the initial allele frequency. Exactly the opposite of masking.

When the deleterious mutation rate is higher than the back-mutation rate (in this case, 40x greater), then we see a totally different dynamic for dominant muts - for some selection coefficients, there is no single equilibrium point - there are three points! Two are stable points, one is unstable

This 'bump' appears to coincide with the point at which the distribution of genotype frequencies is the farthest from Hardy-Weinberg Equilibrium - suggesting another avenue in which the allele frequency is insufficient to describe population dynamics in polyploid populations

First, we found that our results are quite sensitive to mutation rates. If the deleterious mutation rate = back mutation rate, then this is the plot of the equilibrium allele frequency and load of the populations. Nothing surprising here, except there is a 'bump' in the load for the dominant case

A mathematical equation that represents the first order partial differential equation for the change in gamete frequencies for a tetraploid with polysomic inheritance. The expression is long, and is in a tiny text, to highight that it is impossible to intuit anything from this expression alone.

These PDEs are not easy to analytically understand, and are surprisingly much much more complex than the diploid ODEs (it's not just a factor of two difference!). This is just one of the entries of the Jacobian for a tetraploid with polysomic inheritance - don't stare at this for too long, I beg you

In order to probe this question, Sam had to approach it using a nonlinear dynamics framework. Although mut-sel equilibrium is easy in diploids (and can be analyzed using a simple ODE), for polyploids, Sam had to develop a framework to numerically estimate equilbria points with a series of PDEs

And, the only theory work that has been done for any tetraploid is for very idealistic and convenient dominance relationships (completely recessive, perfectly additive) that make the math simpler. We wanted to look at arbitrary dominance relationships, to see if any interesting patterns arose

Although mut-sel equilibria has been estimated for tetraploids with polysomic inheritance, very little theory has been developed for tetraploids with disomic inheritance. We wanted to see if segregation differences changed these mut-sel equilibrium points, or how populations approach these points

Bistable Mutation-Selection Equilibria and Violations of Fisher's Theorem in Tetraploids: Insights from Nonlinear Dynamics Polyploidy and whole genome duplication (WGD) are widespread biological phenomena with substantial cellular, meiotic, and genetic effects. Despite their prevalence and significance across the tree of ...

New Preprint! This one led by Sam Gibbon, an outstanding undergrad I've worked with for the past year. We take a look at whether segregation patterns in tetraploids (disomic v polysomic) affects mut-sel dynamics. We found some surprising results, and some subtleties: www.biorxiv.org/content/10.1...

That's true, with a strong enough drawl, there's no confusing the two. There is a flow cytometer here that is part of the center's shared equipment, it's currently not being used. So it'll migrate down the hallway into my space as soon as I get it ready

Annie Mae might evolve into Anime by accident? Or on purpose, your choice

Oh I'm stealing this for the flow cytometer my lab is inheriting!

Overall, I'm excited to see future broader sampling of high-quality genomes refine the story we've shown here, arguably one of the most complicated polyploid histories in plants. And, there's a ton to be learned about how this ancient polyploid history influences modern biological processes, too!

We found a bunch of traits are associated with this ancient polyploid history, and especially with the rate at which chromosome number (and gene content) is reduced, suggesting that the 'recovery' that lineages much endure from this massive ploidy increase also leads to, e.g., higher speciation rate

And yet, one of the genomes assembled for this paper (Reevisia pubescens) has a totally different polyploid history, even though it is nestled in the middle of this polyploid rich family (it was in the same subfamily as Durio! This is also the reason we suggest a new subfamily designation for Durio)

In this new paper, which has a genome from every subfamily, we discovered that the polyploid history looks to be one giant mess - All three subgenomes of the hexaploid Durio are found in the decaploid cotton, and 2 of the 3 subgenomes in Durio can be found in subfamilies that only have a duplication

When the Durio zibethinus genome came out in 2017, it showed evidence of a whole genome triplication. So not the exact same history as cotton, but we couldn't rule out if that hexaploidy was a stepping stone on the same path that cotton took on its way to become decaploid (10N) or dodecaploid (12N)

When the first diploid cotton genome was published in 2013, there was evidence of an ancient 5- or 6-fold multiplication that wasn't shared by the chocolate genome (one of the early-diverging lineages to the rest of the family). But it wasn't known whether this was a singular or multiple events.

New Paper! In 2019, my first dissertation chapter revealed complicated polyploidies in the cotton family (Malvaceae) but we lacked the tools and genomes to truly understand it. Now with better genomes and improved methods, it's much more complicated than we thought. www.nature.com/articles/s41...

Oh this is really fun!

Logo for the Genomic History Inference Strategies Tournament.

If you're new to demographic history inference from population genomics, try this webapp I created to illustrate how dadi fits bottleneck models to site frequency spectra: ryangutenkunst-dadi-two-epoch.hf.space . It even outputs files for submitting to the GHIST competition! ghi.st

The Conover Lab is open!

I am so excited to join a group of talented, caring, and creative folks at the @danforthcenter.bsky.social, all in service of improving the human condition through plant science.

All you need is read counts for positions you want to call SNPs for. We used GATK for this, but any program that can give allele counts for read mapping works. Feel free to reach out and I'm happy to help debug!

GitHub - dcgerard/updog at rm Flexible Genotyping of Polyploids using Next Generation Sequencing Data - GitHub - dcgerard/updog at rm

We're writing up the paper now, but all of the code is publicly available on the rm branch of the github repo, and works for any even-ploidy population. github.com/dcgerard/upd...