RGRoger Guimerà

Roger Guimerà

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Roger Guimerà has a total of 258 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    From molecular generation to process-relevant performance: AI-driven identification of novel working fluids

    by Manuel Ruiz-Botella, Tiago M. Eusébio, Marta Sales-Pardo, Roger Guimerà, Sabrina Belén Rodriguez-Reartes, Fèlix Llovell

  • Open Access

    Bayesian symbolic regression: automated equation discovery from a physicist’s perspective

    by Roger Guimerà, Marta Sales-Pardo

    Symbolic regression automates the process of learning closed-form mathematical models from data. Standard approaches to symbolic regression, as well as newer deep learning approaches, rely on heuristic model selection criteria, heuristic regularization and heuristic exploration of model space. Here, we discuss the probabilistic approach to symbolic regression, an alternative to such heuristic approaches with direct connections to information theory and statistical physics. We show how the probabilistic approach establishes model plausibility from basic considerations and explicit approximations, and how it provides guarantees of performance that heuristic approaches lack. We also discuss how the probabilistic approach compels us to consider model ensembles, as opposed to single models. This article is part of the discussion meeting issue ‘Symbolic regression in the physical sciences’.


  • Discovering New Working Fluids: An Integrated Computational Framework

    by Tiago Eusébio, M. Ruiz-Botella, Marta Sales-Pardo, Roger Guimerà, Sabrina Belén Rodríguez-Reartes, Fèlix Llovell

  • Top Cited Papers
    Open Access

    ChemEmbed: a deep learning framework for metabolite identification using enhanced MS/MS data and multidimensional molecular embeddings

    by Muhammad Faizan-Khan, Roger Giné, Josep M Badia, Maribel Pérez-Ribera, Manuel Ruiz-Botella, Alexandra Junza, Jordi Capellades, Iván Pérez-López, Shipei Xing, Abubaker Patan, et al.

    Machine learning offers a promising path to annotating the large number of unidentified MS/MS spectra in metabolomics, addressing the limited coverage of current reference spectral libraries. However, existing methods often struggle with the high dimensionality and sparsity of MS/MS spectra and metabolite structures. ChemEmbed tackles these challenges by integrating multidimensional, continuous vector representations of chemical structures with enhanced MS/MS spectra. This enhancement is achieved by merging spectra across multiple collision energies and incorporating calculated neutral losses from 38 472 distinct compounds, providing richer input for a convolutional neural network (CNN). ChemEmbed ranks the correct candidate first in over 42% of cases and within the top five in more than 76% of cases. In external benchmarks such as CASMI 2016 and 2022, ChemEmbed outperforms SIRIUS 6, the current state-of-the-art in computational metabolomics. We applied ChemEmbed to predict structures in the Annotated Recurrent Unidentified Spectra (ARUS) dataset and confirmed 25 previously unidentified compounds. These findings demonstrate ChemEmbed’s potential as a robust, scalable tool for accelerating metabolite identification in untargeted mass spectrometry workflows.


  • Open Access

    Integral Bayesian symbolic regression for optimal discovery of governing equations from scarce and noisy data

    by Oriol Cabanas-Tirapu, Sergio Cobo-Lopez, Savannah E. Sanchez, Forest L. Rohwer, Marta Sales-Pardo, Roger Guimerà

    Understanding how systems evolve over time often requires discovering the differential equations that govern their behavior. Automatically learning these equations from experimental data is challenging when the data are noisy or limited, and existing approaches struggle, in particular, with the estimation of unobserved derivatives. Here, we introduce an integral Bayesian symbolic regression method that learns governing equations directly from raw time-series data, without requiring manual assumptions or error-prone derivative estimation. By sampling the space of symbolic differential equations and evaluating them via numerical integration, our method robustly identifies governing equations even from noisy or scarce data. We show that this approach accurately recovers ground-truth models in synthetic benchmarks, and that it makes quasi-optimal predictions of system dynamics for all noise regimes. Applying this method to bacterial growth experiments across multiple species and substrates, we discover novel growth equations that outperform classical models in accurately capturing all phases of microbial proliferation, including lag, exponential, and saturation. Unlike standard approaches, our method reveals subtle shifts in growth dynamics, such as double ramp-ups or non-canonical transitions, offering a deeper, data-driven understanding of microbial physiology.

  • Open Access

    SingleFrag: a deep learning tool for MS/MS fragment and spectral prediction and metabolite annotation

    by Maribel Pérez-Ribera, Muhammad Faizan-Khan, Roger Giné, Josep M Badia, Alexandra Junza, Oscar Yanes, Marta Sales-Pardo, Roger Guimerà

    Metabolite and small molecule identification via tandem mass spectrometry (MS/MS) involves matching experimental spectra with prerecorded spectra of known compounds. This process is hindered by the current lack of comprehensive reference spectral libraries. To address this gap, we need accurate in silico fragmentation tools for predicting MS/MS spectra of compounds for which empirical spectra do not exist. Here, we present SingleFrag, a novel deep learning tool that predicts individual fragments separately, rather than attempting to predict the entire fragmentation spectrum at once. Our results demonstrate that SingleFrag surpasses state-of-the-art in silico fragmentation tools, providing a powerful method for annotating unknown MS/MS spectra of known compounds. As a proof of concept, we successfully annotate three previously unidentified compounds frequently found in human samples.


  • Open Access

    A collaborative constrained graph diffusion model for the generation of realistic synthetic molecules

    by Manuel Ruiz-Botella, Marta Sales-Pardo, Roger Guimerà

    Developing new molecular compounds is crucial to address pressing challenges, from health to environmental sustainability. However, exploring the molecular space to discover new molecules is difficult due to the vastness of the space. Here we introduce CoCoGraph, a collaborative and constrained graph diffusion model capable of generating molecules that are guaranteed to be chemically valid. Thanks to the constraints built into the model and to the collaborative mechanism, CoCoGraph outperforms state-of-the-art approaches on standard benchmarks while requiring up to an order of magnitude fewer parameters. Analysis of 36 chemical properties also demonstrates that CoCoGraph generates molecules with distributions more closely matching real molecules than current models. Leveraging the model's efficiency, we created a database of 8.2M million synthetically generated molecules and conducted a Turing-like test with organic chemistry experts to further assess the plausibility of the generated molecules, and potential biases and limitations of CoCoGraph.


  • Open Access

    Probabilistic alignment of multiple networks

    by Teresa Lázaro, Roger Guimerà, Marta Sales-Pardo

    The network alignment problem appears in many areas of science and involves finding the optimal mapping between nodes in two or more networks, so as to identify corresponding entities across networks. We propose a probabilistic approach to the problem of network alignment, as well as the corresponding inference algorithms. Unlike heuristic approaches, our approach is transparent in that all model assumptions are explicit; therefore, it is susceptible of being extended and fine tuned by incorporating contextual information that is relevant to a given alignment problem. Also in contrast to current approaches, our method does not yield a single alignment, but rather the whole posterior distribution over alignments. We show that using the whole posterior leads to correct matching of nodes, even in situations where the single most plausible alignment mismatches them. Our approach opens the door to a whole new family of network alignment algorithms, and to their application to problems for which existing methods are perhaps inappropriate.


  • Open Access

    Interplay between children’s cognitive profiles and within-school social interactions is nuanced and differs across ages

    by Eudald Correig-Fraga, Roger Guimerà, Marta Sales-Pardo

    Studies investigating the link between school achievement and social networks have shown that both cognitive and non-cognitive factors are integral to academic success. However, these investigations have predominantly been confined by two limitations: 1) they rarely combine cognitive and social data from the same individuals, and 2) when incorporating social data, it is often unidimensional, focusing only on a single type of relationship among children, such as friendship networks or time spent together. This research builds on prior findings by considering cognitive and social data, including preferences for schoolwork relations, leisure/play relations, and friendships, of nearly 5,000 students from Catalonia (Spain) aged 6 through 15. Our findings indicate that children prefer to interact with those who exhibit similar cognitive profiles, but that their preferences diverge between schoolwork and play-related relations during both primary and secondary school. The diverging preferences of children of older ages suggest a greater understanding of the different purposes and expectations of various social interactions.