Biology subjects

Invernizzi, M.

Publications and source records attributed to Invernizzi, M..

4 recordsLinked to original sources

bioRxiv2026

Transient tertiary structure in intrinsically disordered proteins revealed by multithermal enhanced sampling

Intrinsically disordered proteins populate heterogeneous conformational ensembles that are challenging to characterise. While all-atom molecular dynamics simulations can provide highly detailed insights into dynamic ensembles, achieving sufficient sampling remains difficult. Here, we show that On-the-fly Probability Enhanced Sampling (OPES) in the multithermal ensemble enables efficient generation of atomistic ensembles for disordered peptides and proteins ranging from 15 to 71 residues in length. Using the potential energy as a collective variable, OPES achieves multithermal sampling within a single simulation replica, without replica exchange or extensive parameter tuning. Across multiple systems, OPES yields reweighted ensembles broadly consistent with replica-exchange with solute tempering (REST2) and unbiased simulations, while accelerating convergence and enabling broader exploration of low-population conformational states. Applied to the intrinsically disordered transcriptional coactivator ACTR, OPES reveals transiently structured states in which multiple -helices involved in partner binding fold cooperatively and form tertiary contacts. These rare, partially structured conformations are reversibly sampled during the simulations, consistent with extensive NMR and SAXS data, and could facilitate folding-upon-binding through conformational selection. They may also represent viable targets for drug design or for engineering disordered proteins with customised conformational landscapes. More broadly, our results establish OPES multithermal sampling as a robust and accessible approach for uncovering rare, functionally relevant conformations in intrinsically disordered proteins.

biophysics↗

bioRxiv2025

Advancing Protein Ensemble Predictions Across the Order-Disorder Continuum

While deep learning has transformed structure prediction for ordered proteins, intrinsically disordered proteins remain poorly predicted due to systematic underrepresentation in training data, despite constituting approximately 30% of eukaryotic proteomes. We introduce PeptoneBench, the first benchmark to enable systematic assessment of ensemble generators for both ordered and disordered proteins, integrating diverse experimental observables. Our analysis reveals that existing evaluation metrics exhibit systematic bias toward the structured spectrum of the proteome. Assessment of popular predictors (AlphaFold2, ESMFlow, Boltz2) confirms high accuracy on ordered proteins but shows performance degradation with increasing disorder. We further present PepTron, a flow-matching ensemble generator trained on data augmented with synthetic disordered protein ensembles. On our benchmark PepTron matches BioEmu on disordered regions while maintaining competitive accuracy on ordered protein benchmarks. Our data augmentation approach demonstrates that targeted training strategies can approach the performance of computationally expensive simulation-based methods, establishing a generalizable framework applicable to other protein generative models. All datasets, models, and code are openly available.

biophysics↗

bioRxiv2025

An Arabidopsis receptor-like kinase mediates competitive plant-plant interactions

While competition among plant species is recognized as a major factor affecting crop yield and plant community dynamics, the genetic and molecular mechanisms underlying natural variation of such biotic interactions remain poorly characterized. Here, we report the cloning of a Quantitative Trait Locus previously detected in a Genome-Wide Association Study investigating the competitive response of Arabidopsis thaliana to the presence of the annual bluegrass weed species Poa annua. Using mutant and complementation strategies, we identified ESCAPE 1 (ESC1) as the gene responsible for the natural variation of an escape strategy of A. thaliana in response to the presence of P. annua. ESC1 encodes a proline-rich, extensin-like receptor kinase, also known as PERK13. An RNA-seq experiment revealed that PERK13 functions through different pathways in leaves and roots involving genes associated with responses to biotic and abiotic stresses. Using these RNA-seq together with yeast two-hybrid (Y2H) data, protein-protein interaction network reconstruction revealed two distinct decentralized protein networks in leaves and roots. These findings support the notion of an active response mechanism involved in neighbor detection. The functional validation of ESC1 underlying natural variation in response to competition opens new avenues for a better understanding of the molecular dialogue involved in plant-plant interactions. HighlightIn this study, we identify a receptor-like kinase enabling Arabidopsis thaliana to detect neighboring species through the activation of specific genetic pathways.

plant biology↗

bioRxiv2024

Improving Inverse Folding models at Protein Stability Prediction without additional Training or Data

Deep learning protein sequence models have shown outstanding performance at de novo protein design and variant effect prediction. We substantially improve performance without further training or use of additional experimental data by introducing a second term derived from the models themselves which align outputs for the task of stability prediction. On a task to predict variants which increase protein stability the absolute success probabilities of PO_SCPLOWROTEINC_SCPLOWMPNN and ESMO_SCPLOWIFC_SCPLOW are improved by 11% and 5% respectively. We term these models PO_SCPLOWROTEINC_SCPLOWMPNN-O_SCPLOWDDC_SCPLOWG and ESMO_SCPLOWIFC_SCPLOW-O_SCPLOWDDC_SCPLOWG.

biophysics↗