Biology subjects

Buratto, D.

Publications and source records attributed to Buratto, D..

7 recordsLinked to original sources

bioRxiv2025

ANUBI: A Platform for Affinity Optimization of Proteins and Peptides in Drug Design

The increasing availability of computational power is opening unprecedented opportunities in computational biology and drug design. Computer simulations based on physical models can now reproduce or replace critical biophysical experiments such as binding affinity evaluations in the drug screening process. Here we present ANUBI (ANUBI Nexus for Understanding Binding Interactions), a software package that automates sequence space exploration and binding free energy calculations to optimize protein or peptide drug candidates for improved target binding. Starting from a user-provided molecular model of the drug-target interaction, ANUBI systematically evaluates point mutations in selected regions using Monte Carlo methodology, retaining favorable mutations based on calculated binding affinity differences. We demonstrate that this approach efficiently samples sequence space, generating dozens of optimized variants in timeframes comparable to experimental approaches at substantially reduced cost. As proof of concept, we applied ANUBI to an antibody-antigen complex and a peptide-protein interaction, identifying variants with significantly improved predicted binding energy (approximately 20 kcal/mol, calculated using the MMPBSA method), within 20 days of computation on a single GPU. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=133 SRC="FIGDIR/small/683353v2_ufig1.gif" ALT="Figure 1"> View larger version (35K): org.highwire.dtl.DTLVardef@17fa242org.highwire.dtl.DTLVardef@75cf1dorg.highwire.dtl.DTLVardef@70888aorg.highwire.dtl.DTLVardef@14412d_HPS_FORMAT_FIGEXP M_FIG TOC GRAPHIC C_FIG

biophysics↗

bioRxiv2025

Design of TCR-mimicking binders for pHLA with high potency

The rational design of high-specificity binders to peptide-HLA (pHLA) complexes remains a major challenge in personalized immunotherapy, particularly for shared neoantigens with single-point mutations. To address this, we have developed an integrated framework that combines knowledge-based deep learning with physics-based simulation for the design of highly specific pHLA binders. Applied to p53 R175H-HLA-A*02:01, molecular-dynamics-guided electrostatic filtering yielded nanomolar binders with >10-fold selectivity over wild type. Motivated by the challenge of KRAS G12V-HLA-A*03:01, which lacks electrostatic cues, we implemented message-passing neural network (MPNN)-based optimization strategies to broaden binder design beyond charge-dependent interfaces. When reapplied to p53 R175H, these strategies improved success rates from 2/5 to 5/5, with most candidates exhibiting minimal or no binding to the wild-type complex. Chimera-based and residue-retention approaches within the MPNN optimization further expanded the sequence-structure search space while preserving critical hotspot interactions. This integrated framework thus enables the design of nanomolar, mutation-selective binders to pHLA complexes, advancing next-generation personalized cancer immunotherapies.

biophysics↗

bioRxiv2025

Assessing computational strategies for the evaluation of antibody binding affinities.

Accurate evaluation of binding affinity is critical in drug discovery to identify molecules that bind strongly to their targets while minimizing off-target effects. Although binding affinity calculations are theoretically well-defined, they require exhaustive sampling of configurational space, a step that often requires significant computational resources. In this study, we compare different methods for calculating the binding energy of antibodies targeting a peptide derived from the N-terminus of CXCR2, a GPCR-family protein. Contrary to some previous reports, we find that equilibrium MMPBSA calculations yield better agreement with experimental binding affinities than non-equilibrium potential of mean force evaluations, underscoring the system-dependent performance of these methods. We also observed a modest improvement in accuracy when MMPBSA is combined with replica exchange molecular dynamics, albeit at a significantly higher computational cost. Calculation based on Rosetta force field, instead, produced results that did not correlate with the experimental data. We attribute these findings to two factors, which could limit the applicability of some methodologies that are widely used in the computation the binding energy: the high potency of the antibodies studied and the dominance of hydrophobic interactions between the antibodies and the peptide. Overall, this work provides important insights for optimizing in silico antibody screening strategies.

biophysics↗

bioRxiv2025

Reengineering the antigen optimization process for superior neoantigen vaccine design

Identifying effective neoantigen sequences is essential for enhancing anti-tumor immunity. However, the vast sequence space (>109 possible peptides) and limited accuracy of existing immunogenicity predictors hinder efficient vaccine design for patient-specific human leukocyte antigens (HLAs). We present AlphaVacc, a deep reinforcement learning framework that integrates Monte Carlo Tree Search with a Transformer-based network to optimize antigenic peptides. AlphaVacc outperforms previous generative models in binding-affinity prediction. Experimental validation of 12 AlphaVacc-generated variants of the BING-4 peptide confirmed that 11 showed increased HLA-A*02:01 binding and 7 elicited significant T cell responses. Further testing of 16 single-mutation peptides confirmed computational predictions for 15 candidates, exhibiting a remarkable success rate. AlphaVacc thus provides a powerful tool for designing neoantigen-based cancer vaccines and may accelerate personalized immunotherapies.

bioinformatics↗

bioRxiv2025

In silico design and validation of high-affinity RNA aptamers for SARS-CoV-2 comparable to neutralizing antibodies

Nucleic acid aptamers hold promise for clinical applications, yet understanding their molecular binding mechanisms to target proteins and efficiently optimizing their binding affinities remain challenging. Here, we present CAAMO (Computer-Aided Aptamer Modeling and Optimization), which integrates in silico aptamer design with experimental validation to accelerate the development of aptamer-based RNA therapeutics. Starting from the sequence information of a reported RNA aptamer, Ta, for the SARS-CoV-2 spike protein, our CAAMO method first determines its binding mode with the spike proteins receptor binding domain (RBD) through a multi-strategy computational approach. We then optimize its binding affinity via structure-based rational design. Among the six designed candidates, five were experimentally verified and exhibited enhanced binding affinities compared to the original Ta sequence. Furthermore, we directly compared the binding properties of the RNA aptamers to neutralizing antibodies, and found that the designed aptamer TaG34C demonstrated a comparable binding affinity to the RBD compared to the representative neutralizing antibodies analyzed in this study. This highlights its potential as an alternative to existing COVID-19 antibodies. Our work provides a robust approach for the efficient design of a relatively large number of high-affinity aptamers with complicated topologies. This approach paves the way for the development of aptamer-based RNA diagnostics and therapeutics.

biophysics↗

bioRxiv2024

A Robust Computational Framework for the Optimization of CDK7 Inhibitors as Promising Cancer Therapy

Cyclin-dependent kinase 7 (CDK7) plays a crucial role in cell cycle regulation and transcription, establishing it as a promising target for cancer therapy. Although the covalent inhibitor THZ1 effectively targets CDK7, it presents risks such as short half-life and potential off-target side effects. To address these limitations, we employed a computational workflow integrating virtual screening, molecular dynamics (MD) simulations, and free energy perturbation (FEP) method to design non-covalent CDK7 inhibitors with enhanced selectivity and safety profiles. MD simulations elucidated THZ1s inhibitory mechanism and identified key molecular fragments within its structure. By incorporating fragments from known inhibitors, we introduced extensive non-covalent interactions within the binding pocket, leading to the identification of three novel non-covalent inhibitors with binding affinities comparable to or higher than that of THZ1. Our findings not only introduce promising CDK7 inhibitors but also present a robust computational framework that could accelerate the discovery of kinase-targeted therapeutics.

biophysics↗

bioRxiv2022

RAGE engagement by SARS-CoV-2 enables monocyte infection and underlies COVID-19 severity

The spread of SARS-CoV-2 has fueled the COVID-19 pandemic with its enduring medical and socioeconomic challenges due to subsequent waves and long-term consequences of great concern. Here we charted the molecular basis of COVID-19 pathogenesis, by analysing patients immune response at single-cell resolution across disease course and severity. This approach uncovered cell subpopulation-specific dysregulation in COVID-19 across disease course and severity and identified a severity-associated activation of the receptor for advanced glycation endproduct (RAGE) pathway in monocytes. In vitro experiments confirmed that monocytes bind the SARS-CoV-2 S1-RBD via RAGE and that RAGE-Spike interactions drive monocyte infection. Our results demonstrate that RAGE is a novel functional receptor of SARS-CoV-2 contributing to COVID-19 severity. One-Sentence SummaryMonocyte SARS-CoV-2 infection via the receptor for advanced glycation endproduct triggers severe COVID-19.

immunology↗