iiinova — where your digital self keeps evolving

Reposted by Alvaro Sanchez

Systems Ecology @systems-ecology.bsky.social · 28/09/2026

What is the future of systems ecology in a rapidly changing world? In a new editorial, Deborah M. Gordon argues for studying ecological systems in transformation—and for extending systems thinking across biological scales. Read in Systems Ecology: doi.org/10.67837/2194

doi.org

The future of systems ecology | Systems Ecology

Question: What is the future of systems ecology in a rapidly changing world?

0126

Reposted by Alvaro Sanchez

Ricard Solé @ricardsole.bsky.social · 22/09/2026

We are very happy to announce the launch of Systems Ecology @systems-ecology.bsky.social, a new independent, diamond open-access journal dedicated to advancing systems-level understanding in ecology. systems-ecology.org/index.php/se Check the Founding editorial! systems-ecology.org/index.php/se...

05721

Reposted by Alvaro Sanchez

Systems Ecology @systems-ecology.bsky.social · 21/09/2026

🌱 Systems Ecology is now live! A new independent, researcher-run, diamond OA journal for ecological research through systems science. Free to read. Free to publish. No APCs. We are now open for submissions: systems-ecology.org Please help spread the word!

systems-ecology.org

Systems Ecology

16857

Reposted by Alvaro Sanchez

Guim Aguadé-Gorgorió @guimaguade.bsky.social · 18/09/2026

This looks promising! "...free to read and free to publish, with no subscription barriers for readers and no article processing charges for authors. The journal is run by researchers, for researchers, with the goal of advancing ecology through systems thinking."

0154

Reposted by Alvaro Sanchez

Purushottam Dixit @pdixit.bsky.social · 15/09/2026

1/ Our new paper is out in PRX Life. Physicists' often treat microbiomes as disordered systems with coexistence in very high-dimensional niche spaces. We asked whether empirical data supports this picture, and found the data point somewhere else: journals.aps.org/prxlife/abst...

journals.aps.org

Low-Dimensional Coexistence in Complex Microbial Ecosystems

Analysis of nearly 200 human gut microbiome datasets shows that microbial coexistence is governed by a few effective ecological dimensions, rather than the high-dimensional niches assumed by prevailing theory.

12010

Reposted by Alvaro Sanchez

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

New preprint from the lab: The Latent Simplicity of Microbial Ecological Interactions www.biorxiv.org/content/10.6... We're excited about this one. We find that high-order microbial interactions often obey simple linear laws making microbial communities far more predictable than one might expect

37032

Reposted by Alvaro Sanchez

Filipa Trigo da Roza @filipatr.bsky.social · 03/09/2026

It’s out! 🥹🎉 We’re so incredibly happy and proud to see this work out in the world. And once again, a huge thank you to the great mastermind @jaescudero.bsky.social and all the co-authors; this wouldn’t have been possible without you! doi.org/10.1038/s415... @natmicrobiol.nature.com

doi.org

High-throughput recovery of integron cassettes for gene discovery screens - Nature Microbiology

Integron insertion sites engineered into counterselection markers allow large-scale and high-throughput capture of integron-encoded genes from genetically tractable bacteria or DNA samples.

75129

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

e.g. this is the global epistasis data for a dataset measuring microbiome effect on several life-history traits of Drosophila, by alison gould et al. Frankly I have no idea where I'd start if I wanted to create a CRM for this. Maybe Alison, Will et al have figured it out by now but it's not obvious!

000

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

Anyway, not sure if this answers your question, I get the feeling I am not quite getting it... Perhaps it'd be more constructive and fun to have a zoom and chat it out! 🙂

000

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

We can build F(q) where q is a vector of traits, and those traints are the parameters that appear in the CRM. One could either measure those traits and then try to infer F(q), or leverage the model to figure out what the leading collective modes may be.

100

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

Now, we could instead (perhaps that's what you had in mind?) forget about predictive CRMs and instead formulate trait-based statistical models based on coarse-grained CRMs. Here, instead of the community-function landscape being F(x) where the input variable x is species Presence/absence...

100

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

And this was just to identify the molecule that's being cross-fed. We'd still have to quantitatively model the rates of secretion, uptake etc. I could give you more examples, including ongoing ones in our lab, where we don't know the mechanism of a "simple" function like pyoverdine production.

100

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

For instance, in tis paper we found L. amylovorus stimulated Ethanol production by S. cerevisiae through acetaldehyde cross-feeding. The student leading the project, Felipe, had to combine FBA and lots of lab experiments to confirm this mechanism. www.nature.com/articles/s41...

nature.com

Complex yeast–bacteria interactions affect the yield of industrial ethanol fermentation - Nature Communications

Industrial sugarcane ethanol fermentations are accomplished by a microbial community dominated by S. cerevisiae and co-occurring bacteria. Here, the authors investigate how microbial community composi...

110

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

...we can neglect gene regulation so the per-capita expression of the function is constant over time, ensuring that the function and biomass contributions per species are proportional. In our own experience, identifying the trait responsible for individual interactions can take months of work.

100

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

But you still need to introduce a lot of assumptions, or else do have a lot of knowledge that we do not always have. When per-cell expression of the function is more constitutive, I agree it gets easier and those are the functions that covary strongly with biomass, precisely because...

100

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

Now, this is ONLY IF you want a mechanistic PREDICTIVE model. To build it you need a lot of quantitative information. If you want to create a randomly parameterized model, with the aim of contrasting it to the statistical presence/absence model, things get somewhat easier. The bar is lower...

100

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

another function that models the biotin uptake rate / growth rate of P polymyxa vs biotin concentration. I am not saying you can't do this for a specific system, with a lot of patience you can. But you do need to gather a ton of a priori information for the specific system you want to model!

100

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

And then there's the issue of auxotrophies, e.g. P. polymyxa, a highly competitive strain and a strong amylase secretor, is also a biotin auxotroph. Other strains can cross-feed it but we'd need to then incorporate to the CRM the rate of secretion of biotin / biotin precursors, as well as...

110

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

Current microbial CRMs are not growth-phase aware, as far as I know, so we'd need to figure out how to include that, maybe include a new variable that progresses over time, but then we'd have to figure out how to construct it and parameterize it.

210

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

To do that you must know the vmax and kcat of the enzymes, the secretion rates, and what type they are (e.g. endo vs exoamylases). Besides, many Bacillus strains seem to secrete amylases in a growth-phase dependent manner (stationary vs exptl phase), so you'd need to include that in the model too.

100

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

... are a byproduct of the secreted amylases of all the other species (also, not all types of amylases produce glucose directly). Then keep in mind that expression rate and growth rate are also coupled together. So we'd need equations that keep track of the glucose/maltose as cells grow.

100

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

... so at the minimum you have to introduce a Hill-like function that you'd have to parameterize for each species, but what do you put in the x axis of the Hill? For instance, amylase in many species is under catabolite repression, and glucose & maltose...

100

Alvaro Sanchez @asanchezlab.bsky.social · 30/08/2026

I think it's more complicated than it seems! If you try to model them via CRM, both exoenzyme activity and siderophore accumulation require us to introduce a model of gene expression per cell (i.e. amylase secretion per cell), which requires us to know the regulatory mechanism...

100

Alvaro Sanchez @asanchezlab.bsky.social · 28/08/2026

That being said, I think that creating side by side comparisons where we construct fully sampled community-function landscapes F(x) and compare x as species presence/absence vs x as trait presence/absence (or trait values for the different taxa), for instance, would be very informative & feasible!

000

Alvaro Sanchez @asanchezlab.bsky.social · 28/08/2026

I see no reason why we couldn't accomplish something similar in synthetic microbial communities (less sure about natural ones, though, who knows!). Anyway, I think connecting with mechanism is ultimately the goal anyway, the epistasis-like interactions do give us ways to get there.

100

Alvaro Sanchez @asanchezlab.bsky.social · 28/08/2026

People studying global epistasis in genetics have actually done very cool work connecting the additive variable that is then passed through a non-linear transform to biophysically interpretable, mechanistic variables (e.g. www.pnas.org/doi/10.1073/..., academic.oup.com/mbe/article/...) 7/

pnas.org

Interpretable modeling of genotype–phenotype landscapes with state-of-the-art predictive power | PNAS

Large-scale measurements linking genetic background to biological function have driven a need for models that can incorporate these data for reliab...

120

Alvaro Sanchez @asanchezlab.bsky.social · 28/08/2026

That's essentially what we did for biomass to find out that u_1 (the vector of background averaged additive effects) was the vector of yields (under some limits, that happen to be fulfilled in the 30+ landscapes we constructed, but which may not necessarily be true in other scenarios)... 6/

100

Alvaro Sanchez @asanchezlab.bsky.social · 28/08/2026

... The good news is that u_1 is generally very close to the vector of average additive effects for all functions & communities we've ever looked at. This means that once we have a candidate trait-based model, we can easily calculate this vector from it and check how well it aligns with u_1. ...5/

100

Alvaro Sanchez @asanchezlab.bsky.social · 28/08/2026

...then It's not obvious whether u_1 should be parallel to a simple vector containing an array of single traits like the vector of yields (Y1, Y2,..YN) as opposed to a vector containing non-linear function of combinations of traits (say u_1,i= Yi*e^(Yi/sum_q(Yq))^2))... 4/

110

Alvaro Sanchez @asanchezlab.bsky.social · 28/08/2026

...which I think is likely to be the case a priori for many if not most community functions, it won't be obvious how the leading collective mode depends on vectors of traits (or even which traits matter most). If we don't know what the mechanism explaining the community-function landscape is... 3/

100

Alvaro Sanchez @asanchezlab.bsky.social · 28/08/2026

For instance, extracellular amylolytic activity barely correlates with biomass production. Or, for siderophore production, we don't yet know what traits would explain the first collective mode. Similarly, whenever we don't have a clear a priori mechanistic model of how the function emerges... 2/

200

Alvaro Sanchez @asanchezlab.bsky.social · 28/08/2026

Thanks Seppe! When the function is (or covaries strongly with) biomass & the mechanism of community assembly is dominated by resource-neduated interactions, I think your take is correct. However, for most other functions CRM traits wouldn't necessarily capture the CF Landscape 1/

100

Reposted by Alvaro Sanchez

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

We thus find that community function is predictable not because interactions are absent or weak, but because interactions at different orders are not independent, they're linked by simple linear relationships. What looks like significant intxn complexity can hide a strikingly simple organization.

121

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

This preprint has been a team effort involving all lab members, led by José Camacho-Mateu, @sabelaquiros.bsky.social, @giulioburgio.bsky.social & Andrea Arrabal, together with @alfonsomendana.bsky.social, @migueldiezfdz.bsky.social & Belén Benítez-Domínguez. We've all learned a lot working on it!

120

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

121

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

The most surprising consequence is that presence/absence remained highly predictive of community function even after 10days of population dynamics. Communities moved far from their initial abundances, yet knowing which species were present at the start predicts community biomass at equilibrium!

111

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

This leads to a final hypothesis: if the dominant biomass mode is set by species yield, it should remain stable as species abundances change. We tested this over 10 days of serial passaging. Despite large shifts in abundance, the dominant mode barely moved and stayed aligned with yield.

100

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

This is of course the classic "non-linear transformation of a latent, additive variable" that is the hallmark of global epistasis in genetics, only now demonstrated for ecological communities, and calculated explicitly from independently measurable parameters.

130

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

Importantly, this collapse is not just descriptive. Once the leading mode (u_1) is known, together with its eigenvalue and the additive effects, we can actually calculate the one-dimensional curve that organizes the landscape. With two modes, the same idea gives the full two-dimensional surface.

200

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

This prediction also holds in the experiments. Across all 31 resource environments, the leading biomass mode closely aligned with species yields measured in monoculture. In other words, the dominant organization of community function could be traced back to a simple, measurable species trait.

110

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

Consumer-Resource models gave us a clear prediction: the dominant collective mode that makes community-function landscapes approximately 1-dimensional should be approx. the same as a vector of species traits, formed by the metabolic efficiencies, or yields, of each species.

100

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

This is interesting but provides little ecological insight. What do these collective modes actually represent biologically? For biomass, we turned to consumer-resource models, where species compete for shared resources and their underlying traits are explicitly defined.

100

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

This simple organization has a major consequence: species’ effects on community function do not vary independently across backgrounds. Instead, they rise and fall together. As a result, the full community-function landscape becomes a function of just one or two ecological collective modes.

100

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

Indeed, across all species, higher-order interaction coefficients were strongly related to the corresponding lower-order ones, with slopes predicted by the FEEs. The same pattern held across different functions (e.g. biomass, pyoverdine) and all 31 resource environments.

100

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

To test this prediction, we built complete community-function landscapes containing every combination of five bacterial species, measuring both biomass and pyoverdine production. We then repeated the entire experiment across 31 different resource environments.

110

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

The answer is remarkably simple: if a species follows a good FEE, its pairwise interactions scale with its additive effect, its three-way interactions with the corresponding pairwise ones, and so on across all orders.

110

Alvaro Sanchez @asanchezlab.bsky.social · 21/08/2026

A clue comes from the previous observation of Functional Effect Equations (FEEs): the effect of adding a species often depends linearly on the function of the community it joins. But that effect itself emerges from interactions at all orders. What does a strong FEE tell us about those interactions?

100