Protein fitness optimization is challenged by a vast combinatorial landscape where high-fitness variants are extremely sparse. Many current methods either underperform or require computationally expensive gradient-based sampling. We present CHASE, a framework that repurposes the evolutionary knowledge of pretrained protein language models by compressing their embeddings into a compact latent space. By training a conditional flow-matching model with classifier-free guidance, we enable the direct generation of high-fitness variants without predictor-based guidance during the ODE sampling steps. CHASE achieves state-of-the-art performance on AAV and GFP protein design benchmarks. Finally, we show that bootstrapping with synthetic data can further enhance performance in data-constrained settings.
Improving Protein Optimization with Smoothed Fitness Landscapes
Massachusetts Institute of Technology（麻省理工学院）；Czech Academy of Sciences（捷克科学院）；Czech Technical University in Prague（捷克布拉格技术大学）；University of Chemistry and Technology, Prague（布拉格化学与技术大学）
As protein therapeutics play an important role in almost all medical fields, numerous studies have been conducted on proteins using artificial intelligence. Artificial intelligence has enabled data driven predictions without the need for expensive experiments. Nevertheless, unlike the various molecular fingerprint algorithms that have been developed, protein fingerprint algorithms have rarely been studied. In this study, we proposed the amino acid molecular fingerprints repurposing based protein (AmorProt) fingerprint, a protein sequence representation method that effectively uses the molecula…
Interpreting Protein Language Model Embeddings via Orthogonal Projection for Protein Fitness Prediction
Hasso Plattner Institute, University of Potsdam（波茨坦大学哈索·普拉特纳研究所）；Broad Institute of MIT and Harvard（麻省理工学院及哈佛大学布罗德研究所）；Humboldt University of Berlin（柏林洪堡大学）；Karlsruhe Institute of Technology（卡尔斯鲁厄理工学院）
The integration of machine learning methods into bioinformatics provides particular benefits in identifying how therapeutics effective in one context might have utility in an unknown clinical context or against a novel pathology. We aim to discover the underlying associations between viral proteins and antiviral therapeutics that are effective against them by employing neural network models. Using the National Center for Biotechnology Information virus protein database and the DrugVirus database, which provides a comprehensive report of broad-spectrum antiviral agents (BSAAs) and viruses they …
Potts models and variational autoencoders (VAEs) have recently gained popularity as generative protein sequence models (GPSMs) to explore fitness landscapes and predict the effect of mutations. Despite encouraging results, quantitative characterization and comparison of GPSM-generated probability distributions is still lacking. It is currently unclear whether GPSMs can faithfully reproduce the complex multi-residue mutation patterns observed in natural sequences arising due to epistasis. We develop a set of sequence statistics to assess the "generative capacity" of three GPSMs of recent intere…
Exploring zero-shot structure-based protein fitness prediction
Department of Computer Sciences（计算机科学系）；Department of Biostatistics and Medical Informatics（生物统计学与医学信息学系）；Morgridge Institute for Research（Morgridge研究所）
Robust Optimization in Protein Fitness Landscapes Using Reinforcement Learning in Latent Space
School of Computing, KAIST（韩国科学技术院计算机学院）；Center for Mathematical and Computational Sciences, Institute for Basic Science（基础科学研究院数学与计算科学中心）；Center for Biomolecular and Cellular Structure, Institute for Basic Science（基础科学研究院生物分子与细胞结构中心）；Department of Biological Sciences, KAIST（韩国科学技术院生物科学系）
Linear Fitness Subspace in Protein Language Models Enables Sample-Efficient Directed Evolution
Nanyang Technological University（南洋理工大学）；Sun Yat-sen University（中山大学）；Zhejiang University（浙江大学）；Carnegie Mellon University（卡内基梅隆大学）；Shanghai Institute of Optics and Fine Mechanics（上海光学精密机械研究所）；Zhongnan Hospital, Wuhan University（武汉大学中南医院）；University of Southern California（南加州大学）
Protriever: End-to-End Differentiable Protein Homology Search for Fitness Prediction
Department of Systems Biology, Harvard Medical School（系统生物学系，哈佛医学院）；Department of Computer Science, University of Oxford（计算机科学系，牛津大学）；Department of Computer Science, University of Copenhagen（计算机科学系，哥本哈根大学）；Enzyme research, Novonesis（酶研究，Novonesis）
University of Oxford（牛津大学）；Harvard Medical School（哈佛医学院）；Cohere
The ability to accurately model the fitness landscape of protein sequences is critical to a wide range of applications, from quantifying the effects of human variants on disease likelihood, to predicting immune-escape mutations in viruses and designing novel biotherapeutic proteins. Deep generative models of protein sequences trained on multiple sequence alignments have been the most successful approaches so far to address these tasks. The performance of these methods is however contingent on the availability of sufficiently deep and diverse alignments for reliable training. Their potential sc…
The development of powerful natural language models have increased the ability to learn meaningful representations of protein sequences. In addition, advances in high-throughput mutagenesis, directed evolution, and next-generation sequencing have allowed for the accumulation of large amounts of labeled fitness data. Leveraging these two trends, we introduce Regularized Latent Space Optimization (ReLSO), a deep transformer-based autoencoder which features a highly structured latent space that is trained to jointly generate sequences as well as predict fitness. Through regularized prediction hea…
Evolutionary Profiles for Protein Fitness Prediction
Zhejiang University（浙江大学）；Peking University（北京大学）；East China University of Science and Technology（华东理工大学）；Chengdu University of Information Technology（成都信息工程大学）；Massachusetts Institute of Technology（麻省理工学院）；Zhejiang University of Technology（浙江工业大学）
Multi-Scale Representation Learning for Protein Fitness Prediction
Mila - Québec AI Institute（米拉 - 魁北克人工智能研究所）；Université de Montréal（蒙特利尔大学）；Harvard Medical School（哈佛医学院）；IBM Research（IBM研究院）；Broad Institute（布罗德研究所）；HEC Montréal（蒙特利尔高等商学院）；CIFAR（加拿大高级研究所）
Few-shot Protein Fitness Prediction via In-context Learning and Test-time Training
Harvard Medical School（哈佛医学院）；University of Copenhagen（哥本哈根大学）；Novo Nordisk A/S（诺和诺德公司）；Microsoft Research（微软研究院）；Technical University of Denmark（丹麦技术大学）
ProSpero: Active Learning for Robust Protein Design Beyond Wild-Type Neighborhoods
Institute of AI for Health, Helmholtz Munich（亥姆霍兹慕尼黑人工智能健康研究所）；School of Computation, Information and Technology, Technical University of Munich（慕尼黑技术大学计算、信息与技术学院）；Helmholtz AI（亥姆霍兹人工智能研究所）；Munich Center for Machine Learning（慕尼黑机器学习中心）；Faculty of Mathematics, Informatics and Mechanics, University of Warsaw（华沙大学数学、信息学与力学学院）
Attention-based models trained on protein sequences have demonstrated incredible success at classification and generation tasks relevant for artificial intelligence-driven protein design. However, we lack a sufficient understanding of how very large-scale models and data play a role in effective protein model development. We introduce a suite of protein language models, named ProGen2, that are scaled up to 6.4B parameters and trained on different sequence datasets drawn from over a billion proteins from genomic, metagenomic, and immune repertoire databases. ProGen2 models show state-of-the-art…
Augmenting Biological Fitness Prediction Benchmarks with Landscapes Features from GraphFLA
School of Computer Science and Engineering, University of Electronic Science and Technology of China（电子科技大学计算机科学与工程学院）；Department of Computer Science, University of Exeter（埃克塞特大学计算机科学系）
Designing DNA and protein sequences with improved function has the potential to greatly accelerate synthetic biology. Machine learning models that accurately predict biological fitness from sequence are becoming a powerful tool for molecular design. Activation maximization offers a simple design strategy for differentiable models: one-hot coded sequences are first approximated by a continuous representation which is then iteratively optimized with respect to the predictor oracle by gradient ascent. While elegant, this method suffers from vanishing gradients and may cause predictor pathologies …
Fitness aligned structural modeling enables scalable virtual screening with AuroBind
Shanghai Jiao Tong University（上海交通大学）；Lingang Laboratory（临港实验室）；Sun Yat-sen University（中山大学）；Fudan University（复旦大学）；Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）；Chinese Academy of Sciences（中国科学院）；Massachusetts Institute of Technology（麻省理工学院）；Ningxia Medical University（宁夏医科大学）
Despite being self-supervised, protein language models have shown remarkable performance in fundamental biological tasks such as predicting impact of genetic variation on protein structure and function. The effectiveness of these models on diverse set of tasks suggests that they learn meaningful representations of fitness landscape that can be useful for downstream clinical applications. Here, we interrogate the use of these language models in characterizing known pathogenic mutations in curated, medically actionable genes through an exhaustive search of putative compensatory mutations on each…
Drug-Target Interaction Prediction via Hierarchical Sequential Cross-Attention over Chemical and Protein Language Models
I2A institute, TÉLUQ University（I2A研究所，TÉLUQ大学）；LIO, CRCHUM（LIO，蒙特利尔大学医院研究中心）；LCSI, ESI（LCSI，阿尔及尔国立高等信息学院）；CoMeT UR 24344, Universite de Poitiers（CoMeT UR 24344，普瓦捷大学）；University of Ottawa（渥太华大学）；Lebanese American University（黎巴嫩美国大学）
In-Context Learning can distort the relationship between sequence likelihoods and biological fitness
Computational Biology and Bioinformatics Program（计算生物学与生物信息学项目）；Emeritus Department of Ecology and Evolutionary Biology（退休生态学与进化生物学系）；Department of Evolutionary Biology（进化生物学系）；Hagler Institute for Advanced Studies（哈格勒高级研究 institute）；Department of Physics（物理学系）；Quantitative Biology Institute（定量生物学研究所）
There is considerable interest in predicting the pathogenicity of protein variants in human genes. Due to the sparsity of high quality labels, recent approaches turn to \textit{unsupervised} learning, using Multiple Sequence Alignments (MSAs) to train generative models of natural sequence variation within each gene. These generative models then predict variant likelihood as a proxy to evolutionary fitness. In this work we instead combine this evolutionary principle with pretrained protein language models (LMs), which have already shown promising results in predicting protein structure and func…
Czech Institute of Informatics, Robotics and Cybernetics, Czech Technical University（捷克信息学、机器人学与自动化研究所，捷克技术大学）；Institute of Organic Chemistry and Biochemistry of the Czech Academy of Sciences（捷克科学院有机化学与生物化学研究所）；School of Biological Sciences, Seoul National University（首尔国立大学生物科学学院）；Interdisciplinary Program in Bioinformatics, Seoul National University（首尔国立大学生物信息学跨学科项目）；Institute of Molecular Biology and Genetics, Seoul National University（首尔国立大学分子生物学与遗传学研究所）；Artificial Intelligence Institute, Seoul National University（首尔国立大学人工智能研究所）；CSAIL, Massachusetts Institute of Technology（麻省理工学院计算机科学与人工智能实验室）；Qminers
