From Phase Separation to Chromosome Architecture: Ander Movilla Joins CRM as Beatriu de Pinós Fellow

Ander Movilla Miangolarra has joined the Centre de Recerca Matemàtica with a physicist’s training, a biologist’s curiosity, and a mathematician’s toolkit. His path here traces an unusual trajectory, from studying how droplets form inside cells to understanding how chromosomes fold in three dimensions, but the thread connecting it all is a fascination with how physical principles shape biological behaviour.
Movilla arrives at CRM as a Beatriu de Pinós postdoctoral fellow, a program funded by the Catalan Government’s AGAUR that brings international researchers to Catalonia for three-year research stays.

Working with CRM principal investigator Tomás Alarcón, he’ll join the Mathematical Biology research group at CRM to tackle one of molecular biology’s most intricate puzzles: how the three-dimensional architecture of chromosomes influences which genes get expressed.

A Physicist’s Entry into Biology

Movilla’s background is in physics, he completed his undergraduate degree at the University of the Basque Country, but his PhD at Institut Curie in Paris was already crossing disciplinary boundaries. He worked on reaction-diffusion equations, the mathematical framework that describes how substances spread and interact in space. But these weren’t the textbook versions.

“They weren’t the classic ones that you’d use for solutions where these things don’t happen,” Movilla explains. “They were for cases where, due to different interactions between the solutes, these droplets form.” He was studying liquid-liquid phase separation: how the cytoplasm inside cells spontaneously organises into distinct compartments, like oil separating from water in a vinaigrette. It’s a phenomenon that’s become central to understanding cellular organisation, especially under stress conditions when metabolic components aggregate into visible droplets.

“Describing that mathematically, especially in a way that’s intelligible to a human and not just a thousand-by-thousand matrix that only a computer can understand, that’s also an interesting question from a purely mathematical standpoint.”

The work was computationally driven but biologically motivated. Then, three months into his PhD, reality intervened. “The experimentalists said, well, this can’t actually be done in the end. And so, your project collapses,” he recalls. “You start thinking, okay, now what do we do?” It was an early lesson in research pragmatism. “I think in science, most of the things you try don’t work out,” Movilla says. “Of course, the ones that do work out get published, and everything goes well. But you try many things that don’t work. And I learned that in the first three months of my PhD.”

The Turn Toward Gene Regulation

During his postdoctoral work at the John Innes Centre in Norwich, UK, Movilla shifted his focus to gene regulation and epigenetics. The questions became more explicitly biological: How do cells with identical DNA differentiate into neurons, skin cells, or muscle? How do they maintain their identities across divisions?

The answers involve histones, proteins that bind to DNA and control whether a gene is expressed. Histone modifications form patterns: blocks with one type of modification, others without, and some stretches are empty in the middle. Tomás Alarcón has been working on mathematical models of these patterns for years, and lately he’s been focused on how they relate to chromosome folding. The patterns appear linear when you look at a genome browser, but in the cell nucleus, the DNA is folded in three dimensions, and that matters.

“When you look at the genome browser, as biologists typically do, it shows up as a line,” Movilla explains. “But then in the cell nucleus, it’s all folded.” Regions of the chromosome that are far apart on the linear sequence might be physically adjacent in three-dimensional space, and that proximity matters. “In zones where there are contacts between the chromosomes, it’s much more likely that those histone interactions happen, even if they’re far apart.”

Tomás Alarcón at CRM has been working on mathematical models of histone modifications for years, examining how these chemical patterns relate to chromosome folding. But most of that work has treated the structure as relatively static. Movilla’s project aims to make it dynamic.

Dynamic Networks and Adaptive Systems

When there’s a contact between two parts of the chromosome, you get a network of contacts. If the network is fixed, you know what it is, and how to describe it. But when that network becomes dynamic, the challenge changes entirely. “When that network is no longer static but adaptive, because the histones change due to their own dynamics, which then implies that the contacts change, that network is no longer static,” Movilla says. “And describing that mathematically, especially in a way that’s intelligible to a human and not just a thousand-by-thousand matrix that only a computer can understand, well, that’s also an interesting question from a purely mathematical standpoint.”

This is where the interdisciplinary work becomes essential. Movilla will be working with data from cell lines, clones of pluripotent embryonic cells that can be experimentally differentiated into various cell types. The datasets are massive: expression levels of 20,000 genes at the single-cell level, histone modifications mapped across the genome, and chromosome conformation capture experiments showing three-dimensional contacts. “You end up with gigabytes and gigabytes of data that’s very complicated to work with,” he says. “And I think one of the things that could help us understand it better would be these kinds of mathematical approaches, where you try to extract the most important things from that data.”

One of the aspects of the work Movilla finds compelling is the bidirectional exchange with experimentalists. Mathematical models don’t just analyse data, they generate hypotheses. “At the John Innes Centre, we often tried to encourage information to flow both ways,” he says. “They inform us on how to do the analyses and what would be important in these data. But then our model, based on these analyses, might suggest you could try this experiment, and you might observe this. This model perhaps tells you that this gene could be important.”

Sometimes the predictions are more tentative. “You tell them; I don’t have much confidence, so don’t invest too many resources in this. But I’ll leave it there as an idea.” Other times, given the data and the model’s assumptions, the direction is clearer. “That’s usually what ends up getting done, because experimentalists generally go for what they’re most confident in,” Movilla says.

The work calls for a steady relationship with uncertainty and failure. For Movilla, that feels inseparable from doing research. A project that collapsed three months into his PhD made the point clearly and early. The CRM now welcomes Ander Movilla as a Beatriu de Pinós postdoctoral fellow for the next three years.