Thread by @rafabppinheiro

The networks of interactions between species within biological communities are among the most important and most complex study objects for ecology. Some topologies were observed in different kinds of networks: nestedness and modularity. How can they emerge? Check this thread.

Here l explain some insights of our paper at @ESAEcology. Authors are me, my Ph.D. advisor: @marcoarmello, @CarstenDormann, and Gabriel Felix (Ph.D. candidate at #Unicamp).
@ppgecmvs @ufmg @ESA_org @UniFreiburg
doi.org/10.1002/ecy.27…

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg Nestedness is the tendency of species with less partners to interact with subsets of the partners of species with more partners. See this GIF, with pollinator and plant species. Circles: species; lines: interactions.

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg Modularity is the formation of subgroups of highly connected species (modules). See this GIF, with pollinator and plant species. Circles: species; lines: interactions.

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg We may consider also a compound topology: a modular network, with internally nested modules.
We used bipartite matrices to represent interactions in this study. For instance, rows may be parasite species and columns host species, and filled cells represent interactions.

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg Additionally, we may study interactions in a weighted context, so that beyond the occurrence/non-occurrence of interactions between species, we account for the frequency or intensity of such interactions (gray tones within cells).

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg Some authors suggest that these topologies are common because of a kind of selection on stability. Networks and subnetworks with such topologies are often more stable, thus long lasting and likely to be observed.
onlinelibrary.wiley.com/doi/abs/10.111…

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg Others point to a simpler explanation: topologies may be by-products of the way species are assembled and interact within communities. They result from the ecological and evolutionary processes acting at the species-level.
onlinelibrary.wiley.com/doi/abs/10.111…

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg In this study we derived rules from a hypothesis we proposed some years ago, and built a computational model to simulate the evolution of consumer-resource networks.
Our question: are these rules enough to promote topologies, without invoking selection on stability? 🤷‍♂️🧐

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg The integrative hypothesis of specialization (IHS) is aimed at explaining the adaptive trade-offs on the evolution of consumer-resource interactions. Latter we realized it could be useful to address topology emergence too.

IHS was my Master's project 😊
sciencedirect.com/science/articl…

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg According to the IHS, when resources are similar, there are no adaptive trade-offs faced by consumers in their exploitation. 😎
However, when resources are too heterogeneous, each consumer cannot exploit efficiently all of them. Jack-off-all-trades, master of none!🙄

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg The IHS Model is based on 3 assumptions. None of those include network-level selection.

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg I won't explain here how we implemented these assumptions as a computational model. So, if you want to know more, see Fig 1 and Fig S1: Appendix S1 in our paper.

And let me know if you have any doubts.😜

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg We conducted simulations with different richness of consumers and resources. But, most important, we also varied the heterogeneity of resources.
IHS predicts the heterogeneity of resource defines the intensity of adaptive trade-offs, which in turn drive topology.

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg Here we see the magic happening. A modular network being formed through the evolution of consumers.
Numbers show the round of the simulation.

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg And, here, a nested network.

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg The IHS model produced networks with all the main topologies found in real-world interactions.

TAKE-HOME MESSAGE 1: selection on stability is not necessary to explain network topologies. They may emerge thought simple rules driving the interactions between species.

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg But what have caused different simulations to have different topologies?
If you’re paying attention in this thread you should have a main suspect. Yes, resource heterogeneity. 👏👏

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg We found two main patterns.
Homogeneous resources = highly generalized, highly connected, nested, and non-modular networks.

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg Heterogeneous resources = highly specialized, sparsely connected, non-nested, and modular networks with internally nested modules (compound topology).

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg It was awesome to find these results, because they are coherent with real-world findings. Flores et al. 2011, tested the topology of many phage-bacteria networks, each including a limited phylogenetic diversity, and found high nestedness.
@joshuasweitz pnas.org/cgi/doi/10.107…

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg @joshuasweitz Latter, they tested a highly diverse phage-bacteria network. And, guess what they found.
Yes, a compound topology. Modular at a large scale, nested within modules.
Those are two AWESOME papers. You should read them!
nature.com/articles/ismej…

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg @joshuasweitz As studies of ecological interactions are often focused on a given taxon or functional group, several of the reported nested networks might be, in fact, modules of larger networks with compound topologies.

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg @joshuasweitz TAKE-HOME MESSAGE 2: Interaction networks containing only similar species show patterns that are not observed in heterogeneous networks, as well as a module does not present the topology of the entire network.

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg @joshuasweitz Hope you liked our study! If you want to know more, please check our paper at @ESAEcology. This paper is a chapter of my Ph.D. thesis, mostly developed at the Federal University of Minas Gerais (Brazil), but also as a guest student in the University of Freiburg (Germany).

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg @joshuasweitz If want to contact me, feel free to send DM or visit my website: rbppinheiro.wordpress.com.
And, just in case you liked my work😉, I am currently looking for a PostDoc position.

@ESAEcology @marcoarmello @CarstenDormann @ppgecmvs @ufmg @ESA_org @UniFreiburg @joshuasweitz @threadreaderapp unroll