Biology subjects

Baaden, M.

Publications and source records attributed to Baaden, M..

13 recordsLinked to original sources

bioRxiv2026

Can SMILES be fragmented into a concatenable ordered sequence of retrosynthetically interesting string block ?

Molecules generated by deep learning models are often difficult to synthesize. Their synthetic accessibility can be improved with automated retrosynthetic analysis, which allows for identifying synthons. However, synthons in a SMILES can be scattered throughout the string depending on the path taken through the molecular graph used to generate the SMILES. We tested whether the ensemble of possible SMILES for a molecule can be used to generate a concatenable ordered sequence of string fragments (blocks) from SMILES that match potential synthons obtained through automated retrosynthetic analysis. We found that exhaustively sampling the SMILES space of a molecule improves the coverage of retrosynthetic breaks. We achieved full coverage of retrosynthetic bonds in string form for 85\% of the 1.9 million molecules in the MOSES dataset. Doing so allowed us to test our block SMILES in an unconditional de novo drug design test case with MolGPT and Monte Carlo Tree Search (MCTS). We found that using blocks as an LLM's token did degrade MolGPT performance due to the curse of dimensionality. However, using the SMILES selected by our blocking algorithm with the default SMILES tokenizer improved the reproduction of physico-chemical properties of samples and also improved uniqueness, novelty, and validity. The MCTS outperforms our MolGPT models in terms of validity and novelty. However, samples generated by the MCTS had physico-chemical properties that were further away from the MOSES baseline than the samples produced by molGPT, with an improved distribution of quantitative estimation of drug-likeness (QED).

bioinformatics↗

bioRxiv2026

Deep-learning predictions of biomolecular structures : persistent limitations and new horizons extended by explicit ion addition

The advent of deep learning-driven tools such as AlphaFold has revolutionized the prediction of biomolecular structures, offering unprecedented accuracy and accessibility for proteins, RNA, and their complexes. While these tools have demonstrated remarkable success in benchmarking competitions and enabled experimentalists to generate models with ease, their widespread use has also highlighted persistent challenges. These include difficulties in assessing model confidence, limitations in predicting transmembrane domains, nucleic acids, conformational diversity, and interactions with ions or ligands, as well as the tendency to misfold intrinsically disordered regions (IDRs). In this perspective, we critically evaluate the strengths and limitations of current AI-based structure prediction tools through illustrative examples with a particular emphasis on the impact of explicit ion modelling. We notably report how the explicit addition of a few potassium cations to the prediction of IDRs or G-quadruplexes can trigger massive conformational switches compared to "dry" predictions. On this basis, we suggest modelling sequences both "dry" and in the presence of explicit potassium cations as a simple, practical way to sample alternative conformations and to expose disordered regions that current predictors tend to over-fold. We discuss the importance of reporting confidence metrics in publications to avoid overinterpretation. Furthermore, we address the unique challenges of RNA structure prediction, where data scarcity and structural complexity limit the performance of both classical and deep learning methods. Our analysis underscores the need for continued methodological advancements, integration of complementary computational tools, and expansion of high-quality experimental datasets. TOC Graphic O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=111 SRC="FIGDIR/small/740587v1_ufig1.gif" ALT="Figure 1"> View larger version (16K): org.highwire.dtl.DTLVardef@e950e9org.highwire.dtl.DTLVardef@1bf22b3org.highwire.dtl.DTLVardef@17f5229org.highwire.dtl.DTLVardef@1eb2e9f_HPS_FORMAT_FIGEXP M_FIG C_FIG

biophysics↗

bioRxiv2026

From foldamers to functional pores: a force field for oligourea-based desalination channels

Sea water desalination is a critical solution to global water scarcity, primarily relying on membrane-based technologies and reverse osmosis. Artificial water channels (AWCs) offer a promising solution for next-generation desalination membranes. Very promising channel building blocks for AWC is oligourea foldamers repeats. These biomimetic molecules, composed of unnatural amino acids, self-assemble into protein-like superhelical channels with water-filled pores, making them ideal candidates for selective water transport. In silico studies provide essential support to experimental efforts in designing novel oligoureas. However, such studies require a dedicated force field for now unavailable for these molecules. In this work, we developed a tailored force field for oligoureas by adapting parameters from two established protein force field families: CHARMM (CHARMM36m) and Amber (GAFF2), using their structural similarities to natural amino acids. Our objective was to identify a force field that reliably preserves the structural integrity of oligourea foldamers. We evaluated two distinct oligoureas using molecular dynamics (MD) simulations across increasing system sizes. By comparing simulation results with experimental data, we assessed key structural features, including folding patterns, stability, and pore shapes. Our findings demonstrate that the CHARMM-based force field consistently reproduces experimental observations for both foldamers, outperforming the Amber-based alternative. This newly developed CHARMM-based force field paves the way for further exploration of oligoureas, enabling deeper insights into their stability and efficiency in sea water desalination.

bioengineering↗

bioRxiv2026

Full-Length Structural Modeling of Mitofusins with AlphaFold Reveals a Novel Cross-Type Dimerization and Insights into Oligomerization

Mitochondrial dynamics, involving fission and fusion, are critical for the maintenance, function, distribution, and inheritance of mitochondria, allowing their morphology to adapt to the cells physiological needs. Mitofusins, large GTPase transmembrane proteins, play a role in this process by driving the tethering and fusion of mitochondrial outer membranes. Dysfunction in mitofusins has been associated with neurodegenerative diseases such as Parkinsons, Alzheimers, Huntingtons, and Charcot-Marie-Tooth type 2A, as well as various cancers, where their dysregulated expression influences cell proliferation, invasion, and chemotherapy resistance. Despite their importance, the precise molecular mechanisms underlying mitofusin-mediated fusion remain unclear and require further structural elucidation. In fact, no complete high-resolution structures exist for mitofusins, including their yeast homolog, Fzo1. Here, we generated and analyzed full-length structural models of mitofusins using AlphaFold. Monomeric, dimeric, and tetrameric assemblies were produced, including complexes with fusion partners Ugo1 and SLC25A46. While AlphaFold predicted limited conformational diversity for isolated monomers, structural variability emerged for predicted homo- and hetero-oligomers. Notably, our models reveal a previously undescribed cross-type dimerization mode involving interactions between heptad repeat domains, not reported in current experimental structures. Comparison with recently resolved experimental data further supports the structural relevance of this interface. These full-length models allowed us to propose a new hypothetical mechanism of outer mitochondrial membrane fusion.

bioinformatics↗

bioRxiv2025

Statistical Molecular Interaction Fields: A Fast andInformative Tool for Characterizing RNA and ProteinBinding Pockets

Developing a physical understanding of the interactions between a macro-molecular target and its ligands is a crucial step in structure-based drug design. Although many tools exist to characterize protein-binding pockets in silico, this is not yet the case for RNA, which has only recently been recognized as a suitable target for small ligands. Molecular Interaction Fields (MIF) are a useful tool to characterize the interactions of a given binding pocket. However, classical MIFs heavily rely on the use of probes, which makes their calculation accurate but very specific to the binding partners in question. We develop here a simple version of MIF, that we call Statistical Molecular Interaction Fields (SMIF), based on functional forms inspired by coarse-grained models and parametrized based on PDB structures and previous statistical analysis of the main form of interactions typical of macromolecules, namely hydrogen bonding, stacking, and hydrophobic interactions. We show that these fields, despite their simplicity, are very informative and overall in agreement with pharmacophoric models. Thanks to a carefully optimized code, our calculations are fast and can be performed in bulk on a large set of binding pockets or even on a full macromolecule. As shown in a few representative examples, the latter possibility opens the way to the analysis of systems as large as 20000 to 80000 atoms in relation to the surrounding environment, i.e., a lipidic membrane, a small ligand, or another macromolecular partner, allowing for a detailed visualization of the possible interactions. The complete software and its documentation are available here: https://smiffer.mol3d.tech/

biophysics↗

bioRxiv2025

NOX transmembrane electron transfer is governed by a subtly balanced, self-adjusting charge distribution

NADPH oxidases (NOX) form a family of transmembrane enzymes that catalyze the formation of reactive oxygen species. These are produced thanks to a chain of electron transfers (ET), shuttling electrons from one side of the membrane to the other, using one flavin and two heme cofactors as redox mediators. In this work we investigate the thermodynamics of the electron transfer (ET) between the two hemes contained in the transmembrane domain by means of extensive molecular dynamics simulations. We compare two proteins of the NOX5 isoform, from homo sapiens (hNOX5) and from cylindrospermum stagnale (csNOX5), a cyanobacteria. We study in detail the influence of both the density of negatively charged lipids in the membrane and of the NOX5 aminoacid sequence on the ET thermodynamic balance. The linear response formalism allows us to decompose the variation in free energy into the individual contributions of the system components (protein, membrane, solvent, etc.). We highlight the major compensatory effects of the various components in the global free energy budget in those complex systems. Although the contributions of the protein or the membrane to the ET thermodynamics can be individually strongly modified by a change in the aminoacid sequence or the membrane composition, they are largely compensated by the rest of the heme environment so that the total free energy is always found to be slightly favorable to the electron transfer. To our knowledge, this study is the first to highlight the effect of membrane charge density on inter-heme ET, providing valuable insights into the molecular mechanisms governing ET catalysis in complex membrane systems.

bioinformatics↗

bioRxiv2023

Lys716 in the transmembrane domain of yeast mitofusin Fzo1 modulates anchoring and fusion

Outer mitochondrial membrane (OMM) fusion is an important process for the cell and organism survival, as its dysfunction is linked to neurodegenerative diseases and cancer. The OMM fusion is mediated by members of the dynamin-related protein (DRP) family, named mitofusins. The exact mechanism by which the mitofusins contribute to these diseases, as well as the exact molecular fusion mechanism mediated by mitofusin, remains elusive. We have performed extensive multiscale molecular dynamics simulations using both coarse-grained and all-atom approaches to predict the dimerization of two transmembrane domain (TM) helices of the yeast mitofusin Fzo1. We identify specific residues, such as Lys716, that can modulate dimer stability. Comparison with a previous computational model reveals remarkable differences in helix crossing angles and interfacial contacts. Overall, however, the TM1-TM2 interface appears to be stable in the Martini and CHARMM force fields. Replica-exchange simulations further tune a detailed atomistic model, as confirmed by a remarkable agreement with an independent prediction of the Fzo1-Ugo1 complex by AlphaFold2. Functional implications, including a possible role of Lys716 that could affect membrane interactions during fusion, are suggested and consistent with experiments monitoring mitochondrial respiration of selected Fzo1 mutants.

bioinformatics↗

bioRxiv2023

MDverse: Shedding Light on the Dark Matter of Molecular Dynamics Simulations

The rise of open science and the absence of a global dedicated data repository for molecular dynamics (MD) simulations has led to the accumulation of MD [fi]les in generalist data repositories, constituting the dark matter of MD -- data that is technically accessible, but neither indexed, curated, or easily searchable. Leveraging an original search strategy, we found and indexed about 250,000 [fi]les and 2,000 datasets from Zenodo, Figshare and Open Science Framework. With a focus on [fi]les produced by the Gromacs MD software, we illustrate the potential offered by the mining of publicly available MD data. We identi[fi]ed systems with speci[fi]c molecular composition and were able to characterize essential parameters of MD simulation such as temperature and simulation length, and could identify model resolution, such as all-atom and coarse-grain. Based on this analysis, we inferred metadata to propose a search engine prototype to explore the MD data. To continue in this direction, we call on the community to pursue the effort of sharing MD data, and to report and standardize metadata to reuse this valuable matter.

bioinformatics↗

bioRxiv2021

Between two walls: Modeling the adsorption behavior of β-glucosidase A on bare and SAM-functionalised gold surfaces

The efficient immobilization of enzymes on surfaces remains a complex but central issue in the biomaterials field, which requires us to understand this process at the atomic level. Using a multi-scale approach combining all-atom molecular dynamics and coarse-grain Brownian dynamics simulations, we investigated the adsorption behavior of {beta}-glucosidase A ({beta}GA) on bare and SAM-functionalized gold surfaces. We monitored the enzyme position and orientation during the MD trajectories, and measured the contacts it forms with both surfaces. While the adsorption process has little impact on the protein conformation, it can nonetheless perturb its mechanical properties and catalytic activity. Our results show that compared to the SAM-functionalized surface, the adsorption of {beta}GA on bare gold is more stable, but also less specific, and more likely to disrupt the enzymes function. This observation emphasizes the fact that the structural organization of proteins at the solid interface is a keypoint when designing devices based on enzyme immobilization, as one must find an acceptable stability-activity trade-off. TOC image O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=94 SRC="FIGDIR/small/450859v2_ufig1.gif" ALT="Figure 1"> View larger version (26K): org.highwire.dtl.DTLVardef@d3d0d7org.highwire.dtl.DTLVardef@3403e4org.highwire.dtl.DTLVardef@eb804dorg.highwire.dtl.DTLVardef@13469c2_HPS_FORMAT_FIGEXP M_FIG C_FIG

biophysics↗

bioRxiv2021

A large disordered region confers a wide spanning volume to vertebrate Suppressor of Fused as shown in a trans-species solution study.

Hedgehog (Hh) pathway inhibition by the conserved protein Suppressor of Fused (SuFu) is crucial to vertebrate development. By constrast, SuFu removal has little effect in drosophila. Previous publications showed that the crystal structures of human and drosophila SuFu consist of two ordered domains that are capable of breathing motions upon ligand binding. However, the crystal structure of human SuFu does not give information about 20 N-terminal residues (IDR1) and an eighty-residue-long disordered region (IDR2) in the C-terminus, whose function is important for the pathway repression. These two IDRs are species-dependent. We studied SuFus structure in solution, both with circular dichroism and small angle X-ray scattering, comparing drosophila, zebrafish and human species, to better understand this considerable difference. Our studies show that, in spite of similar crystal structures restricted to ordered domains, drosophila and vertebrate SuFu have very different structures in solution. The IDR2 of vertebrates spans a large area, thus enabling it to reach for partners and be accessible for post-translational modifications. Furthermore, we show that the IDR2 region is highly conserved within phyla but varies in length and sequence, with insects having a shorter disordered region while that of vertebrates is broad and mobile. This major variation may explain the different phenotypes observed upon SuFu removal.

biophysics↗

bioRxiv2020

FAIR sharing of molecular visualization experiences: from pictures in the cloud to collaborative virtual reality exploration in immersive 3D environments

Motivated by the current Covid-19 pandemic that has spurred a substantial flow of structural data we describe how molecular visualization experiences can be used to make these datasets accessible to a broad audience. Using a variety of technology vectors related to the cloud, 3D- and virtual reality gear, we examine how to share curated visualizations of structural biology, modeling and/or bioinformatics datasets for interactive and collaborative exploration. We discuss F.A.I.R. as overarching principle for sharing such visualizations. We provide four initial example scenes related to recent Covid-19 structural data together with a ready-to-use (and share) implementation in the UnityMol software. SynopsisVisualization renders structural molecular data accessible to a broad audience. We describe an approach to share molecular visualization experiences based on FAIR principles. Our workflow is exemplified with recent Covid-19 related data.

scientific communication and education↗

bioRxiv2020

Implicit modeling of the impact of adsorption on solid surfaces for protein mechanics and activity with a coarse-grain representation

Surface immobilized enzymes play a key role in numerous biotechnological applications such as biosensors, biofuel cells or biocatalytic synthesis. As a consequence, the impact of adsorption on the enzyme structure, dynamics and function needs to be understood on the molecular level as it is critical for the improvement of these technologies. With this perspective in mind, we used a theoretical approach for investigating protein local flexibility on the residue scale that couples a simplified protein representation with an elastic network and Brownian Dynamics simulations. The protein adsorption on a solid surface is implicitly modeled via additional external constraints between the residues in contact with the surface. We first performed calculations on a redox enzyme, bilirubin oxidase (BOD) from M. verrucaria, to study the impact of adsorption on its mechanical properties. The resulting rigidity profiles show that, in agreement with the available experimental data, the mechanical variations observed in the adsorbed BOD will depend on its orientation and its anchor residues (i.e. residues that are in contact with the functionalized surface). Additional calculations on ribonuclease A and nitroreductase shed light on how seemingly stable adsorbed enzymes can nonetheless display an important decrease in their catalytic activity resulting from a perturbation of their mechanics and internal dynamics.

biophysics↗

bioRxiv2019

Glutathionylation primes soluble GAPDH for late collapse into insoluble aggregates

Protein aggregation is a complex physiological process, primarily determined by stress-related factors revealing the hidden aggregation propensity of proteins that otherwise are fully soluble. Here we report a mechanism by which glycolytic glyceraldehyde-3-phosphate dehydrogenase of Arabidopsis thaliana (AtGAPC1) is primed to form insoluble aggregates by the glutathionylation of its catalytic cysteine (Cys149). Following a lag phase, glutathionylated AtGAPC1 initiates a self-aggregation process resulting in the formation of branched chains of globular particles made of partially misfolded and totally inactive proteins. GSH molecules within AtGAPC1 active sites are suggested to provide the initial destabilizing signal. The following removal of glutathione by the formation of an alternative disulfide bond between Cys149 and Cys153 reinforces the aggregation process. Besides acting as a protective mechanism against overoxidation, S-glutathionylation of AtGAPC1 triggers an unexpected aggregation pathway with completely different and still unexplored physiological implications.

plant biology↗