Dr.Geneva Answers Our Questions

Some background before we get started here: Dr.Geneva is a Genomics Scientist who researches and has written publications on topics such as Evolution, Genomics, Computational Biology, and Reptiles in general. He used to run a lab at Harvard but now has a lab at Rutgers in Camden, New Jersey that you can learn more about at The Geneva Lab – Evolutionary Genomics @ Rutgers University–Camden

Recently the man has heard of Thrive through somebody I know and wanted to meet me, so I took the opportunity to ask some questions both of my own and from the team (his favorite question was from Thim). The following will be organized by who asked what, with answers being below the question. Fair warning some of my notes were fairly general / summarized. The man can talk.


  • Can you tell me more about speciation and specifically if any algorithms exist?

Currently there is no consensus on what makes a species. In fact, there is even a camp of scientist’ who believe speciation is not real (Geneva disagrees quite a lot). Typically, in scientific simulations speciation may arise organically without explicit implementation, though there are some simulations where this explicit implementation occurs. The most important thing either way is that there needs to be a system of simulating genes and changes at the individual level and not just at the population level. Every new trait first arises in a single individual within a species eventually going from a 1/n representation to a n/n representation (in the new species), this is a concept called fixation. Species are often classed as a rating between the old species and the suspected new species on a 0 - 1 scale with 0 not being able to reproduce with the old species and thus being something completely new, and 1 being the exact same thing. Most times species fall in between meaning there’s a chance of successful reproduction. Whatever we use for speciation, it must be a process that can fail.

At this point Geneva stated that we should really look more into Coalescent Theory for our simulations which is pretty much the above explanation

Another big factor in speciation is Allopatric Speciation (Allopatric speciation - Wikipedia.) This basically just describes physical separation as being a primary culprit for speciation.

  • Can you tell me about Niches and whether or not they are worth explicitly simulating? (I described our current approach, current theory, and past debate)

To summarize, Geneva just completely agreed with our current theorist’ that we shouldn’t explicitly hard code niches, we are pretty much doing things correctly as is, we maybe just need more stuff than food source

I think you guys are already on the right track, I don’t think there should be an explicit class for niches. Instead, you guys should try to simulate the different factors that create different niches such as what something eats, how it behaves, how it interacts with the world around it (this is ecology btw, important to evolution but not the same / not Geneva’s specialty), etc. For creating food webs and describing or observing niches as they arrive, you will need some form of Node Clustering (this is basically Data Science at this point).

  • Can you tell me more about some of the math behind Evolution?

Well, with evolution the only thing that matters in the parent and child (in an OOP sense), everything that happened before that parent is not relevant (which makes things nice to program in these simulators). You guys should look into Markov Chains (Markov Clustering Algorithm. In this post, we describe an… | by Arun Jagota | Towards Data Science.)


  • Can you tell me about the differences between discrete and continuous models in population simulation?

Microbes use completely discrete models, along with very certain animals. However, a continuous model is more representative of most animals. Most of the time we do use discrete though because we only test certain scenarios and conditions, we don’t need the fully continuous simulation every time.

  • How much is the emergence of symmetry dictated by the limitations of DNA? Is there any advantage or biological bias towards symmetry?

DNA certainly may play a role in the emergence of symmetry in animals, which is largely linked to embryotic development. However, there is no limitation that alone makes symmetry so common, microbes for example are almost always asymmetric and during the Cambrian there were many species who exhibited symmetries beyond bilateral symmetry and had very strange body plans.

  • Can we apply time scale dissociation to species population (based on reproductive rate) what would be a good criterion?

Here in my notes all I wrote was that Size and Generation time closely correlate, to where large animals such as Whales take very long to experience lasting changes where microbes can evolve many times in only a month (see Covid-19)


  • Unusual or obscure sources of energy?

Thermosynthesis is the strangest energy source I know of, sounds like you guys already have all the fundamental sources of energy.

  • Viability of Thermosynthesis?

I know there’s lots of research around it. I don’t think it’s been observed but it’s technically possible. You’d be better off asking one of the Microbiologist around here.

  • Mechanisms of Radiosynthesis

At this point Geneva gave me some names of Microbiologist I should maybe talk to at Rutgers

  • Access to papers on the above?

You guys should just reach out to the authors, they’re usually pretty excited to have any readers at all. However, yes if you guys need a paper I can likely get it, just ask.

  • Favorite Anole Fact?

The toes of anoles use special structures called seta to cling to things. this uses interactions between electron clouds to work (Van der Waals force - Wikipedia) and thus requires no effort on the part of the anole. It actually takes more effort to un-cling than to cling, as such, anoles are often found clinging to things even when dead.


  • Why is there more than one kind of algae? (this relates with how in Thrive, autotrophs tend to convergently evolve very harshly and precisely)

This is because of that Allopatric Speciation we were talking about. Some algae as I understand, evolve different pigments and such to function better with different wavelengths of light available at different depths in different places. There’s someone here who studies algae, you should talk to them, they’d be interested and have much better answers.

  • Is something lost in simulations that calculate a whole generation in a single step (discrete) vs those that run continuously? (Geneva’s favorite question)

Good question and yes! Life History ( Life-history theory | Psychology Wiki | Fandom) is totally lost outside of a continuous model. The same fidelity of individual interactions and age differences can make quite a difference and can’t really be done with a discrete model, you guys would be breaking ground if you found an actually accurate way to do that. However, even in this lab we often use discrete models, discrete models aren’t necessarily bad or inaccurate in certain use cases and studies.


  • What’s your favorite dinosaur?

Spinosaurus. Our perception of what kind of dinosaur it was changes literally every few years. It’s really interesting.


Did you ever get to talk to the other people he directed you to?

Not yet, no. They weren’t there at the time

Thank you very much for forwarding our questions to Doctor Geneva, and for laying out their responses in a easily digestible format. I’d be grateful if you could extend my thanks to them for their responses.

I look forward to hearing from those that Geneva pointed your towards should you get the chance to speak with them.

1 Like

Thank you for the answers Nunz!

For the record, what my first question was about was his opinion on continuous vs discrete models as we use them for auto-evo in Thrive. In fact, his answer to Thim’s question is more along the lines of what I expected as an answer. It was still fairly generic however (and also, as Thim put it on Discord, we go a larger step further since we compute over many many generations). Did Geneva express any preference with respect to our “use case and study”? Did he mention any literature on what can’t be done with a discrete model?