Biology subjects

Lupo, U.

Publications and source records attributed to Lupo, U..

6 recordsLinked to original sources

bioRxiv2026

Origin-1: a generative AI platform for de novo antibody design against novel epitopes

0Generative artificial intelligence has advanced antibody discovery, yet de novo design of therapeutic antibodies against targets with "zero-prior" epitopes remains a fundamental challenge. We define "zero-prior" epitopes as target sites lacking structural data from any reported antibody-antigen or protein-protein complex involving the target. Here we present Origin-1, a generative AI platform that overcomes this by integrating epitope-conditioned all-atom structure generation, paired complementarity determining region sequence design, and a specialized co-folding-based scoring protocol to select antibody designs predicted to be high-confidence, specific binders with favorable developability. We evaluated Origin-1 on a panel of ten targets selected to have no available protein-protein complex structures and minimal homology ([≤]60% sequence identity) to proteins with known complexes, creating stringent design conditions. In fewer than one hundred design attempts per target, we identified developable, specific antibodies, validated across multiple biophysical and developability assays, for four targets: COL6A3, AZGP1, CHI3L2, and IL36RA, with functional inhibition demonstrated for IL36RA. Cryogenic electron microscopy confirmed the atomic accuracy of our designs, revealing complexes that closely matched the computational models with high structural fidelity (3.0-3.3 [A] resolution; 0.83-0.91 DockQ). Furthermore, we employed AI-guided affinity maturation to optimize a de novo antibody binder against IL36RA, producing functional antagonists with sub-nanomolar affinities and a top EC50 of 12.3 nM. These results demonstrate a framework for targeting epitopes without structural precedent, expanding the programmable therapeutic antibody landscape.

molecular biology↗

bioRxiv2024

DiffPaSS - High-performance differentiable pairing of protein sequences using soft scores

Identifying interacting partners from two sets of protein sequences has important applications in computational biology. Interacting partners share similarities across species due to their common evolutionary history, and feature correlations in amino acid usage due to the need to maintain complementary interaction interfaces. Thus, the problem of finding interacting pairs can be formulated as searching for a pairing of sequences that maximizes a sequence similarity or a coevolution score. Several methods have been developed to address this problem, applying different approximate optimization methods to different scores. We introduce DiffPaSS, a differentiable framework for flexible, fast, and hyperparameter-free optimization for pairing interacting biological sequences, which can be applied to a wide variety of scores. We apply it to a benchmark prokaryotic dataset, using mutual information and neighbor graph alignment scores. DiffPaSS outperforms existing algorithms for optimizing the same scores. We demonstrate the usefulness of our paired alignments for the prediction of protein complex structure. DiffPaSS does not require sequences to be aligned, and we also apply it to non-aligned sequences from T cell receptors.

bioinformatics↗

bioRxiv2023

Pairing interacting protein sequences using masked language modeling

Predicting which proteins interact together from amino-acid sequences is an important task. We develop a method to pair interacting protein sequences which leverages the power of protein language models trained on multiple sequence alignments, such as MSA Transformer and the EvoFormer module of AlphaFold. We formulate the problem of pairing interacting partners among the paralogs of two protein families in a differentiable way. We introduce a method called DiffPALM that solves it by exploiting the ability of MSA Transformer to fill in masked amino acids in multiple sequence alignments using the surrounding context. MSA Transformer encodes coevolution between functionally or structurally coupled amino acids within protein chains. It also captures inter-chain coevolution, despite being trained on single-chain data. Relying on MSA Transformer without fine-tuning, DiffPALM outperforms existing coevolution-based pairing methods on difficult benchmarks of shallow multiple sequence alignments extracted from ubiquitous prokaryotic protein datasets. It also outperforms an alternative method based on a state-of-the-art protein language model trained on single sequences. Paired alignments of interacting protein sequences are a crucial ingredient of supervised deep learning methods to predict the three-dimensional structure of protein complexes. Starting from sequences paired by DiffPALM substantially improves the structure prediction of some eukaryotic protein complexes by AlphaFold-Multimer. It also achieves competitive performance with using orthology-based pairing. Significance statementDeep learning has brought major advances to the analysis of biological sequences. Self-supervised models, based on approaches from natural language processing and trained on large ensembles of protein sequences, efficiently learn statistical dependence in this data. This includes coevolution patterns between structurally or functionally coupled amino acids, which allows them to capture structural contacts. We propose a method to pair interacting protein sequences which leverages the power of a protein language model trained on multiple sequence alignments. Our method performs well for small datasets that are challenging for existing methods. It can improve structure prediction of protein complexes by supervised methods, which remains more challenging than that of single-chain proteins.

bioinformatics↗

bioRxiv2022

Impact of phylogeny on structural contact inference from protein sequence data

Local and global inference methods have been developed to infer structural contacts from multiple sequence alignments of homologous proteins. They rely on correlations in amino-acid usage at contacting sites. Because homologous proteins share a common ancestry, their sequences also feature phylogenetic correlations, which can impair contact inference. We investigate this effect by generating controlled synthetic data from a minimal model where the importance of contacts and of phylogeny can be tuned. We demonstrate that global inference methods, specifically Potts models, are more resilient to phylogenetic correlations than local methods, based on covariance or mutual information. This holds whether or not phylogenetic corrections are used, and may explain the success of global methods. We analyse the roles of selection strength and of phylogenetic relatedness. We show that sites that mutate early in the phylogeny yield false positive contacts. We consider natural data and realistic synthetic data, and our findings generalise to these cases. Our results highlight the impact of phylogeny on contact prediction from protein sequences and illustrate the interplay between the rich structure of biological data and inference.

bioinformatics↗

bioRxiv2022

Generative power of a protein language model trained on multiple sequence alignments

Computational models starting from large ensembles of evolutionarily related protein sequences capture a representation of protein families and learn constraints associated to protein structure and function. They thus open the possibility for generating novel sequences belonging to protein families. Protein language models trained on multiple sequence alignments, such as MSA Transformer, are highly attractive candidates to this end. We propose and test an iterative method that directly employs the masked language modeling objective to generate sequences using MSA Transformer. We demonstrate that the resulting sequences score as well as natural sequences, for homology, coevolution and structure-based measures. For large protein families, our synthetic sequences have similar or better properties compared to sequences generated by Potts models, including experimentally-validated ones. Moreover, for small protein families, our generation method based on MSA Transformer outperforms Potts models. Our method also more accurately reproduces the higher-order statistics and the distribution of sequences in sequence space of natural data than Potts models. MSA Transformer is thus a strong candidate for protein sequence generation and protein design.

bioinformatics↗

bioRxiv2022

Protein language models trained on multiple sequence alignments learn phylogenetic relationships

Self-supervised neural language models with attention have recently been applied to biological sequence data, advancing structure, function and mutational effect prediction. Some protein language models, including MSA Transformer and AlphaFolds EvoFormer, take multiple sequence alignments (MSAs) of evolutionarily related proteins as inputs. Simple combinations of MSA Transformers row attentions have led to state-of-the-art unsupervised structural contact prediction. We demonstrate that similarly simple, and universal, combinations of MSA Transformers column attentions strongly correlate with Hamming distances between sequences in MSAs. There-fore, MSA-based language models encode detailed phylogenetic relationships. We further show that these models can separate coevolutionary signals encoding functional and structural constraints from phylogenetic correlations reflecting historical contingency. To assess this, we generate synthetic MSAs, either without or with phylogeny, from Potts models trained on natural MSAs. We find that unsupervised contact prediction is substantially more resilient to phylogenetic noise when using MSA Transformer versus inferred Potts models.

bioinformatics↗